Panbo

The N2K WiFi gateway issue, is NMEA stifling innovation?

... written for Panbo by Ben Ellison and posted on Mar 4, 2012
NMEA_2000_gateway_certification_costs_cPanbo.jpg

A few weeks ago I added a comment to an entry about Chetco and DMK Ethernet/WiFi NMEA 2000 gateways that caused some anxiety in the developer community. I reported that "NMEA has essentially served Chetco with a Cease & Desist order regarding its various SeaSmart 'NMEA 2000 compatible' products" and added my opinion that NMEA had a valid case. I'm pleased to report that Chetco has already acted on some of NMEA's demands and it seems likely that their N2K gateways will be certified in due time. And I'm hoping that this entry will clear the air about what happened and why all manufacturers -- and consumers -- should respect the NMEA 2000 certification process...or at least frame the debate...

I'll start with the concern that ActiveCaptain developer Jeff Siegel expressed as part of a recent Forum post. Unfortunately I felt obliged to 'unpublish' the post because it wandered off into strange territory (that I'm still wondering how to handle), but I think I've fairly edited it down to this core complaint:

I have been quite concerned with reports about cease and desist orders given to innovative small companies like Chetco by NMEA and other aggressive acts I've heard about. I've spent my entire career on the innovation side of the technology world. I know how hard it is to start with an idea and an empty basement and create a product with value from absolutely nothing. NMEA is going in the wrong direction if they don't respect that and look for ways to work with that type of innovative entrepreneur.  ~ Jeff Siegel

The most obvious response to Jeff's argument is that no innovator building a product on top of NMEA 2000 is starting with "absolutely nothing"; N2K is, in fact, the result of thousands of hours of largely volunteer labor contributed by members of a marine electronics trade organization that any company can join for a reasonable fee. Moreover the 2000 Standard was conceived from the beginning as copyright-protected intellectual property (IP) with fees for documentation and certification, both to avoid the compatibility issues that grew up around the unprotected 0183 standard and to provide a revenue stream to finance N2K's management and evolution.
   While Panbo has been hosting complaints about NMEA 2000 since 2005, I have to wonder if the critics have paid attention to how much NMEA has lowered the fees and otherwise eased access over the years. The table at the top is specifically about N2K gateways, but I think it's fair to estimate the total cost for a company to join NMEA, get the proper documentation, and certify its first NMEA 2000 product at about $6,000, with the next product at about $1,000. (The major variable in either case is actually testing the product to the standard, but I know that in some cases third party companies do that for about $200.)
   It's taken a long time, but NMEA also instituted a Third Party Gateway (TPG) scheme that makes it's even less expensive for software developers to get N2K Approval when their program is used with a certified gateway. So far only Actisense offers the hardware portion of TPG, and the Approval testing process is not quite finished, but a lot of us are already enjoying the numerous software programs that are working their way toward Approval.
   Not that everything about NMEA 2000 is just ducky! Where, for instance, are the sophisticated alarm messages that I heard were written three years ago?!? I also wish that NMEA had come up with a standard way to encapsulate N2K messages (PGNs) over Ethernet before we all got so interested in mobile marine apps. Which is why I got excited when Chetco announced SeaSmart modules that include open messaging specs so that any software developer could work with them. However, I did wonder about Chetco's status with NMEA, and asked about it when I met the principals in Ft. Lauderdale. I was told that it would cost Chetco nearly $100,000 to get all the SeaSmart modules certified, but that number turned out to be substantially exaggerated (see table above), as did their professed confusion about just how these products fit into the NMEA 2000 rules.
   While it's true that the Chetco gateways contain the same Actisense gateway chip that's in the certified NGT-1, there's more to the story. In fact, there are two ways to incorporate Actisense gateway technology into a hardware product. If you use Actisense's tested and certified printed circuit board, which handles the power aspects of N2K as well as the data, NMEA understandably treats the situation differently than if you just use the Actisense chip, as Chetco does. That's because Chetco had "to surround the OEM Actisense ARM chip on their PCB with all of the necessary NMEA 2000 circuitry in order to maintain the required isolation, plus design, manufacture and EMC test the PCB assembly."
   This may be more detail than anyone wants to know, but the point is that NMEA has good reason to want hardware certification for products like the SeaSmart modules (and note that they created a discount for cases like this where a certified OEM chip is involved). Moreover, Actisense and NMEA told Chetco all this at least a year ago -- I've seen the correspondence.
   So is NMEA stifling innovation by coming down hard on Chetco? I don't think so. Not when it would have cost Chetco only about $15,000 to get 10 SeaSmart models fully tested and certified, and that includes joining the organization where they could be pushing to make their N2K-to-Ethernet scheme the standard. Heck, it's my understanding that getting ten different modules CE tested -- required before selling them in Europe, and elsewhere -- actually would cost $100,000. Shouldn't N2K certification have been near the top of the SeaSmart business plan?
   I also think that trying to avoid NMEA 2000 fees and testing is unfair to all the companies that play by the rules, and I've come to see that there's a difference between uncertified products made by companies that otherwise participate in NMEA and those that don't. I beat on NMEA about this last year, and am still frustrated about certification snafus like daisy chaining, but now I better understand their frustration with "rogue" companies that market NMEA 2000 "compatibility" but don't even join the organization or buy the documentation. I also question the viability of such companies, given the actual costs.
   I'm sorry that Chetco is getting singled out in this entry, but that's the way the story unfolded. The truth is that I could be writing an entry right now about how well the SeaSmart E-Net module feeds NMEA 2000 data to iNavX on my iPad -- and I dare say that Jeff Siegel would have shared my enthusiasm -- but it turned out that the product as it exists right now is illegal and unethical. I don't think that NMEA deserves any blame for this situation, and I'm hoping that a happy ending is in sight.

In the meantime, I'm quite willing to host a debate on this subject, though I am going to enforce some ground rules. For instance, I will delete or edit comments that are critical of specific individuals not in the conversation (that's always true, but it seems to come up regarding subjects like this). I also don't want to see the discussion sidetracked by the argument that NMEA is not an open standard because it's not free. The base definition of an open standard is one that doesn't exclude any manufacturer -- like N2K -- and while "free" is common in the computer world, fees are common in the industrial world that N2K more closely resembles. In fact, I just learned recently that NMEA pays fees for the CanBus intellectual property that N2K is based on. And, please, let's not get into how N2K should have been based on Ethernet in the first place. It wasn't, it's everywhere in marine electronics now, and it integrates with Ethernet in many ways; those are the facts we live with. The question is whether or not NMEA is stifling innovation related to NMEA 2000. I don't see it, but I will listen carefully to any and all comments.

NMEA_2000_software_only_with_certified_gateway_Approval_costs_cPanbo.jpg

Comments

As a member of the boating industry, I know how hard it is to meet the fees of various bodies that want to charge a fee for their certification and compliance. You make light of it, and in these days where profit margins are very slim, large fees do in fact stifle innovation and product development. Your "reasonable fee" could quite well be a burden to some small or start up company. We, ourselves, stopped a development product, because of all the compliance costs and the downturn in the boating industry. We nearly had to close our doors.
An alternative might be NMEA certifing the product for free and charging, say, 0.5% on every unit sold to collect their monies.
snip "The base definition of an open standard is one that doesn't exclude any manufacturer -- like N2K -- and while "free" is common in the computer world, fees are common in the industrial world that N2K more closely resembles." It certainly isn't an open standard, with NMEA claiming IP rights, certification and fees, no one I know of would call that an "open standard". Fees affect a business bottom line, and any fee that is too high to the extent that it excludes a business/product, does not lend itself to being an "open standard" and stifles development.

Posted by: Paul at March 4, 2012 6:58 PM

Paul, I certainly didn't mean to "make light" of any fee a small business has to contend with. I have filed a Schedule C (or was supposed to have ;-) almost every one of my 45 working years and have only once held a salaried job (not quite full time, six years); I know very well what it's like to own a micro business.

But that doesn't mean I can just build a product on the work of others without consequences. Nor could I, say, expect the Fort Lauderdale International Boat Show to give me a free booth to show it off. How could NMEA and FLIBS do what they do for the boating industry without contributions from the industry? In fact, a software developer, and NMEA member, recently told me that NMEA should charge higher fees for NMEA 2000 if it would help accelerate the standard's development!

I asked commenters not to wander off into arguments about what an open standard is, but since you have, perhaps you'll be willing to fully ponder the Wikipedia entry on the subject. It's a complex subject but I stand by my "base definition":

http://en.wikipedia.org/wiki/Open_standard

But, please, let's be more specific. If someone is going to argue that the NMEA 2000 fees are stifling, don't they also have to explain why a business that can't handle a $6,000 one-time expense for its first hardware product based on a hard-to-create multi-manufacturer standard -- and $1,000 for subsequent products -- deserves any attention from consumers?

Posted by: Ben in reply to Paul at March 4, 2012 7:42 PM

I recognise that the term "open standard" has various definitions, I include here the New Zealand take on open standards.

New Zealand official interoperability framework definition

The E-Government Interoperability Framework (e-GIF) [17] defines open standard as royalty free according to the following text:

While a universally agreed definition of "open standards" is unlikely to be resolved in the near future, the e-GIF accepts that a definition of “open standards” needs to recognise a continuum that ranges from closed to open, and encompasses varying degrees of "openness." To guide readers in this respect, the e-GIF endorses "open standards" that exhibit the following properties:

Be accessible to everyone free of charge: no discrimination between users, and no payment or other considerations should be required as a condition to use the standard.
Remain accessible to everyone free of charge: owners should renounce their options, if any, to limit access to the standard at a later date.
Be documented in all its details: all aspects of the standard should be transparent and documented, and both access to and use of the documentation should be free.

The e-GIF performs the same function in e-government as the Road Code does on the highways. Driving would be excessively costly, inefficient, and ineffective if road rules had to be agreed each time one vehicle encountered another.

It appears New Zealand may have a different take on "open standard" than your wikipedia. I still hold to my premise that NMEA is not an "open standard" and stifles development within that field of the industry. Isn't this "stifling" what the debate is about? Not whether a business can handle a $6000 fee, by which I'm sure there may be some.

Posted by: Paul at March 4, 2012 8:12 PM

PS, Paul. I think Chetco is also advocating for an N2K licensing scheme instead of up-front fees. But I'm skeptical. Wouldn't that pile more bureaucratic chores on NMEA's plate, while providing no extra funding for maintaining and extending the N2K standard?

It's worth noting that NMEA is a tiny organization, with only about five salaried employees, I think. I know that some folks out there like to picture NMEA as some sort of Darth Vadar monopolistic consortium, but I'm pretty sure that those folks have never, say, attended an NMEA annual conference and actually broken bread with the "despots" they like to vilify.

Remember that NMEA is an organization founded by independent installer/dealers, often "mom & pop" operations. A real problem they need to address -- and worthy of another entry -- is bringing in other major players in marine electronics, like boat builders and some big box retailers.

By the way, it would be useful if knowledgeable marine electronics manufacturers chimed in with information about the costs of other necessary standards compliance. It seems like NMEA gets singled out more than it deserves.

Posted by: Ben in reply to Paul at March 4, 2012 8:19 PM

NMEA2000 is the poster child for How to Frustrate the Adoption of Technology.

The anti-competitive behavior of NMEA is only 40 years behind the times. If you remember when this all started, I tried to explain that "Connectathons" were the *only* approach which has been shown to reliably develop interoperability and that all the "certification testing" in the world would not accomplish it. That opinion was based on long, painful experience building networks. And now what do they have? They have an NMEA2000 Connectathon! So why do they still require "certification testing"? Why don't they just require participation in Connectathons? Because having to pony up the money up front prevents the riff-raff from showing up with inexpensive hardware and systems that will upset their economic fantasy. It's happened everywhere else in electronics and communication. Why should marine electronics be any different? Because there is an organization who has created a "standard" designed to prevent it from happening. Not even Microsoft could hold back the tide forever. I personally hope that Chetco and the independent software producers file an anti-trust complaint with the FTC and the EC Competition Commission. There is no possible reason do demand fealty from the software people working downstream from a NMEA2000 gateway. The next thing you know they will claim they are owed a license fee from everyone who makes a web browser who can look at the HTML version of a web page showing data collected by NMEA2000.

it's time for NMEA to get with the current reality voluntarily before someone helps them, painfully.
It will happen eventually. The question is how much damage will be done before they get a clue.

Posted by: Mike at March 4, 2012 9:40 PM

I am a patent attorney who works with many start-ups and individuals hoping to start a company, so I know quite a bit about the difficulties start-ups face when when it comes to costs. I also know a bit about the reasons for protecting the intellectual property of others. Ben's points are well thought out and correct in my view. First, the work of others must be respected. If we don't, next time that work will not be available. This is nothing new as evidenced by the fact that the bases for both copyright and patent protection are set out in the US constitution. Second, six thousand or so for certification is not really a lot of money. Compare that cost to costs for manufacturing, advertising, professional services (legal, accounting, etc.), etc., and any business owner would agree it is not a lot of money. Would it be better for small businesses if it was free? Of course. But it would be great if your server, oscilloscope, rent, etc. were free too. Unfortunately, people want to get paid for their efforts and using the NMEA 2000 standard uses the efforts of many. Lastly, as Ben also pointed out, if you are making a product and you can't come up with six grand, should you really expect anyone to connect your product to their network where there average display is around $6,000 and that is just a small part of the system?

Posted by: Guest at March 4, 2012 10:49 PM

Well said, Guest!

And inasmuch as I was actually hoping that Mike would show up, I must be a masochist ;-) But let's have at it:

Mike, if I recall correctly you once worked at Bell Labs, the very birthplace of Linux, which is both the poster child for free and open source software and probably also the reason that so many techies don't seem to understand how open industrial standards work. Can you name an analog to Bell Labs in the marine electronics industry? Do you have any experience with CanBus standards, CE certifications, etc.? For instance, what sort of fees and standards testing were required for the Ossa Powerlite Propulsion system you were involved with?

I believe that you are currently a venture capitalist. Instead of advising Chetco to go to the Government about NMEA, why not help them? Seriously, I think it's an under-financed company with great ideas. But also seriously, if you actually were an investor in Chetco, what would you do? Bleed resources fighting NMEA or spend a (very) few bucks to get right with the standards-making body and then advance your product?

And, yes, I do remember your bleeting about how important "Connectathons" were. NMEA has actually been holding "ConnectFests" for a very long time, but I don't think they are a substitute for lab testing of software and hardware compatibility.

Mike, the bottom line 2012 is that NMEA 2000 has become pervasive in marine electronics, despite all your doubts. As imperfect as it may be -- I'm certainly not arguing that -- it's still the best news out there for people trying to build systems that will endure and grow.

And there are plenty of small companies that have not found the NMEA "bar to entry" difficult. Consider these guys, for instance: http://goo.gl/xL7RY Are you ready to tell Ron and John -- "just two guys from Appleton, Wisconsin" -- how NMEA is stifling their innovations? I think they'll be surprised.

Posted by: Ben at March 5, 2012 12:14 AM

To Guest,
To be sure, IP rights should be paid for, but when we have a world wide industry standard, one that is purported to be an open standard, the fees should not be excessive. Especially since the purpose of such a standard such as N2K is to unite the industry and prevent the issues Ben and others have written about, with the NMEA 0183 standard. Unfortunately, N2K, does have issues, and in part due to the closed nature of the standard. Not everyone got their voice heard for the standard. And as such there were/are gaps.
What is clear to me, if the industry wants to implode with excessive fees for certifying a N2K product, it won't be long before another group gets together with a major industry leader and create a new standard. Lets call it " Marine Industry Electronics Interface" standard. A standard perhaps created to be "freeware". And they may base it on any number of existing data methods. Make it free to get your product certified and suddenly you have not only a competing standard, but one that may potentially catch on and leave N2K behind.
There seems too many points of view as to whether the fees are affordable for the development of a product or not.
I know from my own experience and with talking with others, fees charged for certification of products do add to the cost, and in one case stifled our development of a product. We decided not to proceed in the end. We will not be the first or last business to stop development due to compliance costs.
Good for Ron and John though. But since they are paying in US dollars, don't have the added burden of an exchange rate that other developing countries are forced to contend with.
In the end, business will decide to produce a product and get it certified or not. How will we know when a business decides not to proceed with a product development and for what reasons? We will never know unless that business writes about it. I for one look forward to the next standard, hopefully one that doesn't have the same issues N2K does and is truly an open standard, free for all, and a certification process that is not burdensome.

Posted by: Paul in reply to Guest at March 5, 2012 2:26 AM

I have been working at an Internet company for the past 12 years (where standards and IP) are a big deal. I am also a marine enthusiast.

Looking at this from all sides, I come down slightly on the Chetco side. (Disclosure: I have installed a Chetco Ethernet gateway on my boat.)

I am not that familiar with the NMEA organization, but from what I have seen, I have found them to be either MIA, or wrong headed. Ben mentioned the daisy chaining issue. I have personal experience with the consequences of some of the requirements for certification. A big one was when NMEA decided that terminating resistors could not be installed in devices. (As Airmar used to have in their PB200 and Furuno still has in the FI instruments.)

Airmar wanted to be certified and responded by removing the resistor from the device and adding a terminating resistor option in their long cables. This means when there is a problem, the cable has to be replaced. Perhaps not an issue at deck level, but replacing the cable running up a mast is not fun.

I understand the theory behind this particular decision, but marine installations are exercises in practice, not theory.


In terms of MIA, the best example is the glacial publication of new PGNs. (Think AIS, for example.) This forces product companies--who are not the fastest moving in this area either--to use propriety PGNs. And of course, propriety means that they won't play nice with products from other manufacturers.

I like what Chetco is doing because they are pushing the envelope. They are working well with iNavx. I have iNavx working on my iPad and iPhone. Both display charts and instruments, over the Internet showing boat status. Recently, AIS data was added. (This is the only example of NMEA 2000 AIS data coming into portable devices I am aware of.)

In some ways, what Chetco is a bit rough around the edges. The web pages displayed by the device could use some work. (Or I can go to work on them myself.)

But the good thing is that Chetco and iNavx are both moving fast and innovating.

From an IP perspective, NMEA has a point. Having been involved in more than a few start-ups, a per unit licensing fee has to be the way to go.

It is possible to protect IP and innovate at the same time. :)

Posted by: Jeremy at March 5, 2012 6:37 AM

Jeremy, I share your frustrations about N2K and I think I've written about all of them. But I don't understand what you mean by being "on the Chetco side". Do you really think it's OK just to use NMEA's IP and ignore copyrights, fees, and the feelings of all the people that actually built the standard?

Let's be clear that NMEA did not try to slow down or change Chetco's Gateways. Assuming that Chetco designed and built its PCBs properly, the box you have on your boat right now could be a tested and certified product. Neither was anything but the fairly minor business expenses detailed above stopping Chetco from becoming a full participant in NMEA. It's not a closed shop.

Posted by: Ben at March 5, 2012 7:46 AM

Hi Ben -

I've been involved with electronic and software standards, since, hmmm.... I think it was around Ben Franklin's time. Actually I was one of the co-founders of the MIDI standards organization for electronic music that is a much more widely used than NMEA will ever be and I was a representative on the Motion Picture Experts Group (MPEG) committee that worked to develop the standards on every digitally recorded movie or TV show that you watch. There were fee structures involved with these organizations and politics were rampant in the development of the standards, but the thing that seems different to me was that the goal was always to promote compliance as a desirable goal for a manufacturer and work to make interoperability easy to do.

I don't see that with NMEA. It really feels, and it's only a feeling, that it's more about making money for the organization.

Secondly, NMEA 2000 does kind of suck. We've been working with it a lot and it just doesn't scale well. The concept of hundreds of devices on a NMEA bus without crashing it is unfathomable. You may think that "hundreds" is just crazy but it's not long ago that tens or hundred of computers in a car, or smart lightbulbs, or any of the other networked devices we live with were crazy. With the power control systems that we build, I would be very concerned if our own communication between breakers and controls and motors and batteries could be crashed by some other NMEA device. Building even modest sized networks or NMEA-2000 devices without crashing the bus is a challenge.

Instead of focusing on owning all the protocol layers, NMEA would have been smarter separating them and making the higher levels of the protocol -- the levels with the actual information much more accessible and open. They could license the cabling and electrical standards separately, and use their trademark Draconian approach for these levels. Then, like CANbus, the information could sit on a low-cost Ethernet, twisted-pair, or (like Hybrid Propulsion does) they could communicate right over the power lines. Although the NMEA-2000 connectors are robust, the connectors and cabling are way too expensive for low cost devices.

I don't really want to get into a re-hash of the pros and cons of the NMEA standard since it's all been done before. But I'm arguing that any standard has to evolve rapidly to meet new requirements or it will be ignored and die.

On Ethernet with TCPIP, billions of nodes manage to coexist with this kind of approach. With MIDI, stadiums full of light arrays and full orchestras are managed by an evolving standard that makes things easier for the manufacturer and isn't seen as a roadblock to development.

The whole boat industry can only exist as a parasite on the technology of the big guys. We have to follow, use and adapt the engineering that's being done for us by the auto, communications, and computer industries. The problem for boats isn't assigning packet standards, it's keeping cables from falling apart and corroding.

My soapbox is starting to sag. I better get off it now before it collapses.

So long for now.

- Bill

Posted by: Bill Southworth at March 5, 2012 9:29 AM

I can see this from both sides. From the new world open tech ideas NMEA seems to cause a problematic hurdle to innovation. On the other side those who play in fully open enviroments know they are often not as user friendly as closed ones. (thou that depends on the user personnaly I will go with open source (free) if I can and deal with the issues usually learning something in the process). Now as it has been said nothing in this world is free and there should be something fed to NMEA for their work I think it should be much simplier then it is. I worked for a company working on monitoring systems for a while 10 years ago and when we researced creating a NMEA product, NMEA's answers to cost were never straight forward. Also if the device did not fit with their current model they seemed unready to provide the answer. To me it would seem it should be a simple matter of one cost to order the standard (much like ABYC) then another set cost to certify the product and I belive The total cost should not run more the $2000 to me this would allow small start ups to buy the standard, test in the field, then order the cert testing once they are able. I seem to remeber in the past some bigs guns in the marine world did not certify their devices not because of cost but because NMEA would not allow the flexibility they required. To me this is a good sign there is something flawed with the NMEA thinking and process. I would assume $2000 per product would allow the small guy in while still creating funding (smaller price should lead to more products.) for advancement. I do appreciate Ben that you broke out the costs above as I don't think NMEA has done a good job of explainaing these and should really work on that in order to get companies to come around to their thinking. When I called NMEA to discuss this I was met with "depends we will have to get back to you" That 100k to certify is a very common rumor in the marine world now so I imagine if NMEA laid out the costs on their website they may be met with more enthusiasm.

Posted by: Guest Chris at March 5, 2012 10:20 AM

This is an interesting debate - and one I've been involved in for years, but in a different industry.

In order for a networking standard (or protocol) to be open, it must be defined enough to be built to, it must be robust enough to handle the innovation introduced to it by developers, it must be developed on it's own in order to keep up with the same innovation it spawns, and products that claim adherence must be tested against the standard to which they were built.

As a result of these demands, this technology, along the same lines as the products that use it, must go through a product development process. Which means the product must be managed. Product management means innovation by it's own right, but also means documentation and administration. And like it or not - this adds cost to what was birthed as a free or near free "open" standard.

I don't know a lot about NMEA as an organization - but if it's like other open networking standards used in other industries (I'm familiar with those used in commercial.industrial building automation) - it probably follows a pretty common "staffing" structure. They'll have an Executive Director to oversee the organization - and that person will draw a salary. They'll have admin help - and that staff will also draw a salary. And they'll have technical staff - staff that will oversee compliance of submitted product and that will assist developers with their questions as their own product development cycles unfold. And those folks will also draw a salary.

When you couple these costs, along with the costs of a physical location (office), printing costs, Internet (website) costs, mailing costs etc. - you have the makings of a small business. aAd as with any business - it needs revenue to make it sustainable.

In the world of building automation, we too had "rogue" developers - and while their ideas were usually good, and their products added to the mix of existing products, they were almost always problematic for those that tried to integrate them into a network of compliant products. In the end and when this happened, it was the system integrators (installers) or the customers that suffered. And the companies that tried to short-cut the process, were usually onto other things in short order leaving the product you purchased - an orphan.

Bottom line - for a open standard or networking protocol to be strong enough that industry can rely on it to foundation their own product, it must be treated like a product in it's own right. And while the organization may choose not to be profitable at this - it still requires revenue to be sustainable.

Product development carries a cost. At an initial cost of approx $6K to have the entire networking portion of any new product defined and developed (and able to be purchased as a sub-assembly)is a bargain. And to continue that accessibility at the cost of approx $1000 per subsequent product is peanuts. Compared to the industry I came from - this is incredibly good value. And when compared against the total cost of any product development exercise that results in a product I want to rely on - it's insignificant.

For me - this is simple. Life's too short to take a chance on non-certified product. If a company wants to circumvent the NMEA process - so be-it. but another company is right here that has not circumvented the process and their stuff is going to work. If that costs me (the consumer) a few more bucks - I'll gladly pay the money. I have no interest in placing the safety of me or my passengers in the hands of a company that's not willing to spend $6K on product development and simple compliance.


Posted by: Nikko at March 5, 2012 10:54 AM

Software developers (see "numerous software programs" link above) who use N2K gateways (see such as Actisense (or Actisense embedded in the Chetco SeaSmart) have a much lower barrier to entry. In good faith, we simply purchase the N2K PGN specification from NMEA. My recollection the price was around US$600 for the specification PDF file (which also included NMEA membership for a year) in early 2010. We act in good faith when we support a N2K gateway and adhere to the NMEA PGN specification.

Maybe I missed it, but I didn't see in this post if the "DMK Box" was properly certified. I note the "DMK Box" remains for sale with a web site that states "NMEA 2000 inputs".

It has always been my goal with all the apps (Mac, iPhone, iPad) to work with open documented standards and not lock the user into one vendor. I will continue to so in regards to live vessel data.

Hopefully Chetco will get the ducks in order with NMEA and can move on from this speed bump that has no impact on the actual functionality of their SeaSmart network or the iNavX software.

Posted by: GPSNavX at March 5, 2012 11:01 AM

I don't think any system is perfect, but a standard is probably better than none and if an organization takes the time and resources to develop an acceptable standard than I believe they should be compensated. I also believe that rather than stifling product development, have a standard will actually be of great benefit. Rather than a free for all with every manufacturer doing their own thing, I now have the option to use the Autopilot I like with the MFD of my choice and the depth sounder I prefer and not be limited to a single manufacturer. This benefits all of those that I choose. Without these options, in the past, I have passed up certain manufacturers because parts of their systems were not to my liking. As a marine technician for many years, the most common complaint I have heard is the inability to integrate different pieces of equipment, noting the "proprietary" nature of manufacturers. I think that those with complaints about the fees and licensing involved in NMEA certainly have the opportunity to go out and develop a standard that is acceptable to all of the manufacturers and give it away if they like. But I would lay my bets down that once those standards have been developed, they will decide that compensation might be a good idea given the time and resources involved in such and undertaking. Chuck

Posted by: Capn. Chuck at March 5, 2012 11:20 AM

Ben (and Anonymous Esq.) your arguements are compelling. I am now with NMEA on this one. However, it appears that this confrontation spun out of control when Egos clashed. That is a warning sign to potential investors, and should be to potential customers too.

Posted by: Anonymous at March 5, 2012 11:27 AM

I agree with Paul's comments and think he may be onto something with the royalty structure. In any event I think having the total fee structure back-end loaded would help to encourage innovation. I don’t know if that is the goal of NMEA, but if it is, back-end loading the fees would certainly help. A royalty structure also helps address the issue of a $10 app download vs. a $3,500 plotter.

How about making the documentation a free download? It’s still copyrighted and the owner maintains full control (“by downloading this documentation you are agreeing to the terms of the license agreement…”). Same goes for the testing/certification software. Allow and encourage a product to be labeled something like “NMEA 2000 beta” with no obstacles; end users would then know what they are getting into. If you want to move to “NMEA 2000 certified” you have to submit the results of the testing suite and pay a royalty.

The problem I see with the current setup is that a developer has to spend $5k just to learn enough to see if he/she wants to play with it. The next stage is playing with it. Then you start to think about the applications and possibilities. Then maybe you throw something up to see if it sticks. If it sticks, paying royalties and reasonable fees isn’t a problem.

Posted by: Robert in reply to Paul at March 5, 2012 12:09 PM

Developers that want to develop a commercial product based on NMEA, in my opinion, should become a member of the NMEA and buy the necessary documentation from them. The only thing holding back commercial developers is cost -- there is a fee associated with becoming a NMEA member, and for the documentation as well. One could argue that the fee structure is wrong (as it favors big companies over small companies) but I think that is something for commercial developers to discuss with the NMEA.

However, someone who wants to develop an open source program or utility cannot become a member and obtain the documentation, as members of the NMEA are not allowed to divulge the exact content of the PGNs to non members. Thus for the open source community there is no alternative but to reverse engineer the data structures from scratch.

This now is what I have done -- just by looking at the public documentation that the NMEA has provided and then trying to make sense of the bits you find on an actual live NMEA 2000 bus. Such reverse engineering is perfectly legal, and in my opinion moral as well as long as I do not claim that it is 'official' NMEA information. As of today, my efforts are licensed under a Creative Commons Attribution NonCommercial-ShareAlike license, so that they cannot be used for commercial products.

Obviously the NMEA does what it feels is best for their organization, and they have the right to exercise their copyrights.

However, I do think it is a mistake not to have a process or program in place that allows open source development on top of the NMEA 2000 protocol.

Posted by: Kees at March 5, 2012 12:47 PM

The attorney is patently wrong. (Sorry, I couldn't resist!)

But seriously… attorneys don't innovate, they protect. Read the Steve Jobs biography if you are unclear as to how real innovation happens. The process by which Apple was started (zero money, zero business plan, a hacked together product that stuck, etc.) actually isn’t unusual. Google was started the same way. Facebook. Twitter. Having been a former VC and software developer prior to that, I can say that it isn’t a fluke that they all started out messy, without a business plan, and merely an attitude of “this might be cool, let’s try it.” Personally, I love the process but hate the failure rate. I’ve tried to find a better alternative for 15 years – but haven’t.

A trade association might be inclined to protect the interests of its current members, and not necessarily be motivated to encourage a raft of new competitors that upset the existing industry structure. This is rational behavior.

I’d encourage a scrappy engineer to smack together a more modern spec and throw it up on a wiki for comment. Hopefully it will get some free press on this site. It will either stick or it won’t.


Posted by: Robert in reply to Guest at March 5, 2012 12:55 PM

GPSNavX, When I asked DMK about NMEA status back in early January, the reply was:

"NMEA certification is not in our immediate plans because it presents some hassles we might want to avoid. Still it is something we could be open to in the future and having the NMEA 2000 spec would help our development greatly."

While I think their box has real potential, I don't think that's a sound development and/or business plan. I don't know if NMEA has contacted them, but I wouldn't be surprised. We haven't really discussed it but it seems obvious that a "rogue" gateway would be particularly galling to the NMEA membership, because it's not just a new and different sensor that's trying to be "N2K compatible" without going through the certification process; it's all about leveraging the N2K standard. (Except that DMK also does NMEA 0183 and SeaTalk.)

At any rate, I started thinking about the ethics of this before I found out that NMEA was concerned, and I don't always catch the obvious ;-). And please note again that I've seen no sign whatsoever that NMEA is trying to prevent these products from existing, as long as they abide by the rules. The topic is whether those rules are truly stifling and I appreciate those who've stuck to it.

Incidentally, there's a fairly clear explanation of the general N2K hardware certification process at NMEA.org under the Standards/2000 tab, and lots of pricing in the store, but I guess you have contact the organization to get the nuances of gateway certification and software approval. My tables, by the way, were vetted by NMEA.

Posted by: Ben in reply to GPSNavX at March 5, 2012 1:04 PM

NMEA has never contacted me in regards to support of N2K in any of my apps. I am hopeful that is because I did legitimately buy the N2K documentation and that application certification should go through the gateway provider (i.e. Actisense).

I believe RosePoint Coastal Explorer also supports the Chetco SeaSmart protocol.

I found Chetco very easy to work with as they are always responsive to my emails.

Posted by: GPSNavX at March 5, 2012 1:15 PM

Ben and contributors:

Thank you for your comments about the NMEA standards and costs. A few points:
- the word 'safety' has not appeared yet in this conversation. Integrity and safe operation of a marine electronics network will not be achieved using ad hoc untested products. Improper electronics interfacing/interconnection can create fires, pull down navigation systems and risk lives at sea. Glitches on a desktop system on land are less hazardous.

- NMEA standards are created by unpaid volunteers, investing tens of thousands of man-hours, and they represent virtually every established marine electronics manufacturer and 100s of technical dealers. NMEA 2000 was copyrighted on their behalf.

- On behalf of its members, NMEA will pursue individuals and companies who are illegally using NMEA 2000 intellectual property and assist them to achieve certification. It is immoral to flout IP law for profit based upon personal opinions.

- NMEA 0183 is a worldwide serial communications standard that has been available free to use for 30 years. Developers are advised to purchase the NMEA 0183 standard, to ensure correct implementation.

- NMEA is a non-profit association and the staff operate on a modest administration budget, of which approx. 20% is derived from standards fees.

- For a current status of the standards updates (done by volunteers in spare time) see the link http://www.nmea.org/Assets/20120301%20nmea%20standards%20update%20march%202012.pdf

- Developers are invited to contact the NMEA national office at info@NMEA.org for any questions about fees.

Posted by: Bruce Angus at March 5, 2012 1:27 PM

I couldn’t agree with you more Bill. Why not hack together an open data level spec and toss it up on a wiki or whatever for comment? XML format? Put the important subset of N2k data in a human readable format? Allow for rapid community extension? There is no reason why a $20 smart LED bulb shouldn’t be able to be controlled from a $2k plotter with zero proprietary hardware between them.

Posted by: Robert in reply to Bill Southworth at March 5, 2012 1:50 PM

Hello,

I don't know about Chetco, but one thing with DMK is that they don't send anything to the NMEA 2000 / CAN bus. They don't even claim an address.

It just has connections for CAN high and low, although you can also power it from the NMEA 2000 power cables. It would be nicer if they added a NMEA 2000 connector to the box though.

Unless there are electrical ramifications, it should have no impact on a NMEA 2000 system.

The NMEA 2000 effort is admirable, but suffers from two main problems. As mentioned above, it is rapidly reaching its limits. Also, the standard is maintained and developed by people who are not at the leading edge. My big bugbear example was always that the default speed update frequency is 1Hz, which is essentially useless on a sailboat

Nick

Posted by: Nick at March 5, 2012 3:00 PM

when I see comments like this one:

..... Integrity and safe operation of a marine electronics network will not be achieved using ad hoc untested products. Improper electronics interfacing/interconnection can create fires, pull down navigation systems and risk lives at sea.....

I shake my head. I understand the passion behind these comments (and applaud it), and I understand the drive for certification. In fact if you scroll up - I supported it through experience and analogy. And I still do.

But blanket comments such as this are irresponsible and will only serve to alienate folks and developers who you really want to enroll in the program.

As one who's fought the fight before (but in a different industry), and as one who believes strongly in the foundation on which NMEA 2000 is premised - might I suggest a bit of a "honey" approach to fence sitters and those currently developing "ad hoc" product, and a bit less emphasis on the "vinegar" approach?

Posted by: Nikko at March 5, 2012 3:01 PM

Angus -

I think your comments about IP protection show just where NMEA has gone astray. With other standards bodies and standards I've been involved with or use, I haven't heard the same level of griping about the unfair fees or restrictive practices of the standards organization.

In most cases, the standards organization is or uses an "expert group" of technology types from its membership. The actual work to draft the standard is done on a volunteer or donated basis. The cost of distribution of the standards is modest and should cover costs. In this age of pdf files, the cost should be negligible. Certification is an area where the standards organization can charge for services but should be voluntary on the part of the manufacturer in order to get a "seal of approval".

If you think that NMEA's paramilitary enforcement of its IP makes the NMEA 2000 network any safer, you are just dreaming. Try hooking up a random collection of say twenty NMEA-2000 compliant devices and you have a pretty good chance that you'll get at least a bus crash, some data overruns, and a few devices that will interfere with each other. Admittedly, much of this is due to the very nature of the CAN bus. To try to lead developers to think that because the devices are certified, it's going to be ok for a device that needs monitoring and control for safe operation to be controlled on a shared NMEA-2000 network is totally irresponsible.

- Bill

Posted by: Bill Southworth at March 5, 2012 3:02 PM


Bell Labs had nothing to do with Linux, but everything to do with UNIX.

Posted by: Brian M in reply to Ben at March 5, 2012 3:13 PM

NMEA does not have a case. When someone says that something is compatible, that does not infer that it is certified by any agency. Users can chose to seek a certified product or accept the product on its claims. What the developer is missing out on is the possibility to build and conform to a standard and all that goes with it. I'd fight them, since it sets a bad precedence. Open source, means its open. Open source does not infer that it also must be certified. This line will continue to get blurred the more we go software and firmware and less on hardware.

Posted by: jehines3 at March 5, 2012 4:19 PM

First off, as co-founder and CTO of Chetco Digital Instruments, I must state we have not and do not engage in illegal activity. We manufacture a wide variety of marine related products. Those that utilize NMEA 2000 interfaces are only a small portion of our product line. We have never stated any of our products are yet NMEA 2000 certified, only that they do meet the requirements as stated. Those customers that wish to purchase a NEMA 2000 certified solution can simply add an external Actisense NGT gateway if they choose that option.

The whole aspect of NMEA 2000 Certification is complex and at times not well understood. While it may be debated as to how much about the process we should have known, it is not that simple when you must pay first to find out.

All of our products that contain NMEA 2000 interface capability, including SeaSmart.net are based on a OEM relationship with Antisense, creators of the NMEA 2000 certified NGT. We felt that using an already certified gateway design would be the most cost effective way to market. The gateway provided all the direct interface with the NMEA 2000 bus and our products would use the now isolated data on the user end. All product development was done exclusively with Actisense Monitoring tools and third-party equipment such as chart plotters and other publically available devices. At no time did we have access to any NMEA restricted or copyrighted documents. As stated in previous posts, this was all done using common and legal practices.

Recently the NMEA has stated displeasure with our mention of NMEA 2000, stating we have somehow engaged in illegal and immoral activity. We disagree. However, Checto Digital Instruments is in the process of obtaining formal NMEA 2000 certification in order to offer its customers confidence they are purchasing the best of quality products.

I have to disagree with some posts that claim that a $10K investment in unnecessary certification is “peanuts” For small innovative hardware companies $10K in cash is very significant. Not only does new product development require talented personal, lab resources, and inventory – it all has to be paid out of sales revenues. For a new product like SeaSmart, funding has to come from other product lines with about $100K in gross sales to come up with the cash for NMEA certs alone. We have to look very carefully as to actual value we get out of such expenditures. Paying this sort of upfront fee is certainly stifling to small companies and keeps them from being part of NMEA. We have always maintained and suggested that a per unit fee or revenue based fee system be a much more fair and equitable approach. As a former Apple Developer in the 80’s – a major key to their success was the active recruitment and support it gave all its “partners”. With NMEA, you need to pay $675 just to see a document that outlines what the certification requirements/tests are! No matter if you can pass them or not. I was involved with Firewire and USB development and this sort of practice would have never worked.

Even more disturbing is now a new NMEA requirement aimed at any Software developer that wants to display any NMEA 2000 data. A new Appendix H has been added to force “Apps” to only use “Certified” Third Party Gateways (TPG) in order to even mention NMEA 2000 or PGN and that they must meet a very restrictive set of tests totally unrelated to their function in order to pass, Its $725 (combined) for non-members.

“..All Third Party Application Gateways must be Certified by NMEA and applications using them must be approved for each Certified Gateway they can be used with. The Third Party Gateway must adhere to the regular NMEA 2000® Certification process and fees. “

I am not aware of any device that can now meet the requirements of Appendix H and therefor assume no “App” can list or claim any official " NMEA compatibility ". This is a case where an overly aggressive stance has completely killed an emerging market. With no “official” platform to develop for, “App” developers are discouraged. NMEA should have used the few already established certified Gateways as a starting point for Appendix H. CDI products could have therefore complied and the spirit of the spec been full filled. The problem here is not the $300 fee but the fact there seems to be no way to implement the spec. While it has been difficult to pin down NMEA on this new wrinkle, it seems as though with a wink, one can just claim “…compatible with company’s XYZ product” and get by with it for now.

On a note about published fees. Total costs for members would be close to $8,400 for a single product with cert tools. All fees shown are almost double for non-members and one must purchase the Certification tools $2,400 to $4,000 if you can’t find someone else to do it for you. Well, it might be asked why would anyone pay an extra $6K-$8K ($15K total) instead of the membership fees? I suspect some simply do not want to be forced to join an organization they do not agree with and I would certainly not want to demand that of our developers.

CE fees are high too but many aspects of CE allow for self-certification. So, for example if a new product uses a WiFi radio that is CE certified and a Gateway that is CE certified, self-certification can be done with minimal expense. The manufacture of the WiFi radio may have to pay a higher fee but that encourages the OEMs to embed their products without the added CE costs. There is no force to purchase docs/spec and no membership fees. Just because a WiFi radio is put in a different color box does not force recertification and extended fees as is the case with NMEA. More importantly, CE certification is optional and without fear if one does not choose to do so.

Joe Burke
CTO
Chetco Digital Instruments
www.chetcodigital.com

Posted by: oregonjoe at March 5, 2012 6:37 PM

Ben asked "is NMEA stifling innovation"? Undoubtably so for small upstarts and small products. If you want to offer a certified product for 100 bucks, you better sell a bunch to a micro market to make it worthwhile. Is there a better way? I'd like to think so, but standards are expensive to test and validate. I suggest that the standards should be published for anyone to read, including us end users at no cost. If a manufacturer wants to claim certified compliance, then by all means have fees for that. But there does need to be a way for innovaters to work within the published standard at some level less than "certified".

The bigger companies still get much more say in what the standard is going to be. That should be sufficient.

I hope the next generation of NMEA 2K follows along something like this. We need it today! How long do we have to wait for a standard Radar data interface?

Bob

Posted by: Bob at March 5, 2012 7:22 PM

I normally lurk quietly in the background, but I feel the need to comment. I am a hobbyist, and I am tinkering with the NMEA 2000 network on my boat. I will never produce any commercial product. I'd like to purchase the NMEA2000 specs, but at over $2K, that is just not an option. Instead, I am reverse engineering things.. If anything breaks, it's my own fault.
As for open standards.. NMEA 2000 isn't new.. It's DeviceNet with a new name..It's ISO J1939 with new PGNs.. You can buy NMEA 2000 connectors, tee's, and cable from a marine supplier for a lot of money, or you can get the identical supplies for half price from your local electronic supplier. Why the enourmous cost difference?
Someone commented that certification is required to prevent the sky from falling and boats from catching fire. The NMEA2000 standard is supposed to achieve interoperability and safety. If that is true, why do we have vendors with proprietary connectors and proprietary PGNs? Clearly, the standard is an epic failure when it comes to compliance because innovative people have to work around it, instead of within it. Please don't get me wrong, I don't think this is a technical issue. I think the problem is with the implementation.
I don't believe NMEA has any interest in innovation.If you want innovation, look at the internet. Little or no license fees, interoperable standards without a lot of certifications, and protocols are the openly published. Most companies adopt standards voluntarily without being forced. There is a model that works. It's been around for decades, and it promotes innovation.

OK.. I'm done ranting..

Posted by: Anonymous at March 5, 2012 7:46 PM

An interesting result of NMEA’s “pay for docs” policy and not share with non-members is how does one know exactly what a “NMEA Certified” product is? Is there anywhere on the web site that states exactly what the certification tests are that is available to non-members?

As a member we are not allowed to disclose what the tests are or the pass/fail criteria but how does this help the consumer?

We can put a sticker on the unit but how does the consumer know what it means without access to the specs and test procedures?

Joe Burke
CTO
Chetco Digital Instruments
www.chetcodigital.com

Posted by: oregonjoe at March 5, 2012 7:57 PM

Dinosaurs. will. die.

Posted by: Ramrod at March 5, 2012 7:59 PM

Hey Joe, All that pretzel twisting and you still didn't address the central question of ethics. Isn't it true that up until recently your company has been marketing products hugely leveraged to NMEA 2000, yet not paying a dime into the standards-making body nor contributing a minute toward advancing the standard? Is that fair to the membership of NMEA, which pays the fees and donates the time, and which definitely includes companies no larger than yours?

Posted by: Ben in reply to oregonjoe at March 5, 2012 9:01 PM

Is the NMEA 2000 standard rather too pricey for small developers and start-ups? Yes. It's not much as a percentage of the cost of R&D on a new product, but to a struggling start-up, every dollar counts- especially on the software side where the dollar amounts involved are so tiny at first.

Are we picking on the NMEA too much? Also yes. Standards organizations do have costs they have to cover. True, it costs zero dollars to distribute a copy of the standard, but it took a fair bit of time and money to put it together, and the NMEA's staff do need to get paid somehow. A per-unit-sold royalty would be prohibitively complex for such a small organization to administer, which is probably why they do it the way they do.

"NMEA 2000" is a registered trademark. The certification costs include paying for the right to use the trademark on that product. IANAL, but I'm pretty sure it's illegal to use the NMEA 2000 trademark in connection with non-certified products.

From the customer's perspective, the tests and the text of the standard don't matter: it boils down to "Will it work, yes/no". If he adds a widget to his network and it crashes the network, the answer is "no" and he exchanges it for a different one. If your product's not N2K certified, there's probably a greater risk that weird edge cases will cause it to hiccup, i.e. a greater risk that customers will be throwing the things back at you with RMA slips attached. Likewise, being certified does not immunize you against weird edge cases (or against having to take responsibility for your device doing something dumb), but it does improve the odds that such issues will be caught and fixed before release.

Re. proprietary connectors: IMHO, this was a dumb mistake that slipped through and got locked in before anyone had a chance to tell Ray and Sim to quit playing games. If you don't like it, simply don't buy from the companies involved (and write to their head offices to let them know why they lost the sale).

Posted by: Matt Marsh at March 5, 2012 9:10 PM

A favorite recently-learned quote came from departed (and disgraced) Governor Lester Maddox who once observed that "the problem with conditions in Georgia prisons is the quality of the inmates." Do you catch my drift?

The reason I tried to set ground rules at the top of this entry is that any discussion of NMEA and/or N2K almost always seems to draw especially critical comments. But that does not mean that the larger readership (over 3,500 today!) share these opinions. In fact, if you read the comments critically you may notice a paucity of actual NMEA type electronics developers as well as the astute consumers trying to build an advanced and reliable electronics system. There are definitely some interesting comments here, but having read almost every Panbo comment for some six years, I can tell you that the malarkey level is higher than normal.

Cherished Vonnegut now: "So it goes"

Posted by: Ben at March 5, 2012 9:38 PM

Ben

Our NMEA 2000 products are leveraged off of the Actisense NGT which we pay a significant COG for. Actisense undertook the task of developing and marketing a certified NMEA 2000 gateway. They recoup some of the costs paid to NMEA by resealing products to other companies such as CDI. All communication takes place on the “public” side of the gateway which is the purpose of a gateway and was certified by NMEA for that purpose. There is nothing “unethical” about any company using this approach to build up new and innovative solutions. All information used for development was legally obtained from Actisense and nothing was required from NMEA.

We pay Actisense for the right to use their technology. No, ethics problems here.

To state CDI has not paid a “dime” is misleading. We do not obtain our access to NMEA 2000 data for free. We pay for each and every node. The problem is that NMEA does not collect any additional revenue if a company sells one product or a million. When Actisense sells to us they get to keep above and beyond what they paid to NMEA. I do think this is the problem but not one CDI created. One has to admit this a great business model for Actisense and a poor one for NMEA.

The argument that NMEA is somehow wronged by this is like saying a software developer has to pay Microsoft just because their Apps run in windows 7. Or that anyone creating a Web Site has to pay AT&T because it goes over phone lines. Yes, the new products are built on existing technologies.

Keep in mind that most of the internet technologies used today are based on another Standard Organization called the IEEE yet we are not being ask to join to use a router or pay a doc fee to connect to a Hot Spot. Nor are the manufactures of these products required to pay. Standards committees create standards to promote market growth. The NMEA members should be quite happy to have new applications that promote sales of their own products. We all do not need to be members to contribute.

What has created the stir with NMEA is that our interface is repackaged into a smaller footprint and intended to be used in multiple applications. They felt that even though the Software and firmware we purchased was exactly the same as the certified NGT, the electrical isolation may be different (it is not) and therefore the whole product now requires recertification from the ground up, taking into no account that 90% of the certification spec deals with the Software/Protocol interaction already covered by our OEM purchase. There are no exceptions for products “based on” or “Licensed from”.

In order to assure customers, CDI has undertaken separate certification so there may be no doubt as to long term availability. There has been no design changes made for this purpose. It is purely a marketing decision.

In essence, we are paying twice for the same product.

Now, NMEA also wants any “App” developer to pay to use any “gateway” products. If I were to question ethics, I would start there. The NMEA want to collect a fee from everyone no matter how far removed from the bus they become. First it’s the connector/cable companies, then the gateways, then routers/access points, and now the app developers. In essence, they want to tax the data the NMEA 2000 is generating. No wonder, product developers like Active Captain are becoming disappointed as their very existence depends on availability of data.

It appears that NMEA discovered a loophole in allowing Gateways to become “certified” in the first place and now is trying to close the barn doors. The new restrictions on Appendix H are confusing and poorly implemented to the point that no one can offer a product. It is my hope that we can work with NMEA to revise these new rules to meet sensible and practical implementations and give developers a clear path to delivering what customers want.


Joe Burke
CTO
Chetco Digital Instruments
www.chetcodigital.com

Posted by: oregonjoe at March 5, 2012 10:56 PM

Sigh,

NMEA is a "trades association" that just happens to own valuable IP. With regards to standards, a trade organization is close to the ultimate in "herding cats". To me its remarkable that NMEA 2000 is as robust as it is, and its clearly a credit to those who have managed the standard in this environment.

This is very different from typical software standards creation and management, many of which originated with the sole intent of providing a competitive advantage to the sponsor. Its also very different from component standards say for semi-conductors to enable the truly bulk sale of goods. In either example the costs of maintaining the standards are absorbed by the principles and passed on in infinitesimal increments to the end consumer.

So who gets to underwrite creating and maintaining a standard within NMEA? To me it would seem that the competitive advantage is greatest for product developers and manufacturers and therefore they should underwrite this effort.

Finally, while Joe Burke is very passionate about the cost of doing business, the costs have to be paid to stay in business. While NMEA is a non-profit, this doesn't mean they don't balance the books. And despite the arguments put forth to the contrary, NMEA costs are probably quite fair. It would be wonderful if there were thousands of developers and manufacturers supporting the standard. The individual cost could be quite low. Nonetheless, at best, there are a hundred or so......

Posted by: Don Joyce at March 6, 2012 2:08 AM

Ben,

I would like to comment on "In fact, if you read the comments critically you may notice a paucity of actual NMEA type electronics developers as well as the astute consumers trying to build an advanced and reliable electronics system."

I don't understand why you say that. Almost every comment here, and there are a lot, seems to originate from either set -- manufacturers, developers or savvy users.

"I can tell you that the malarkey level is higher than normal."

I don't understand why you say this. Every post I've read in this thread is on-topic, at least half restrained, and not abusive. Some are expressed better than others, sure. But malarkey?

Is it that most people seem to think that yes the NMEA is stifling innovation?


How can a low cost or 'gratis' commercial app developer justify $1369 with $625 recurring every year? In a market where he will, in all likelyhood, only sell a few hundred copies? The entire point is not that someone will come up with the killer app that will be a runaway success (and sell maybe 10.000?) The point is that this way you won't get the tens or hundreds of small app developers, each trying something different, because of these "modest" and "fair" barriers to entry.

And to get back to my point of open source: these developers cannot participate at all, period, since the NMEA does not allow republication of the data in any format.

Posted by: Kees at March 6, 2012 5:00 AM

To keep on point, costs are only part of the stifling criteria. Some think its high, some low, and I have expressed my view.

However, the other factors come into play as well. The restrictive sublicensing, locked-up documentation, control over form and function, all add to the mix.

The NMEA wants everyone to pay but how far removed from the bus should they be allowed to reach?

Now, if we were making end-user products like chart plotters, it may not be such an issue. Everything is locked up and controlled by the manufacture and users really do not need to know anything about NMEA 2000. No one else can develop applications or add-ons so it does not matter.

When we looked at initial costs for a single product, the math did not work. So, we had to try and amortize it over many other products to try to make it work. WiFi, Ethernet, Displays, data sources and so on. NMEA wants to collect a fee on every variant even when it is all really the same. That is stifling. Now, they also want to collect fees from anyone who uses our product of its intended purpose. If we provide a Web Server and someone creates a new web page, will they want them to pay too?

The stifling part comes in when smaller/cheaper enabler products are developed (gateways) specifically to allow more free access to the data. Here the NMEA needs to rethink its policy to conform with modern times. Data access is the way of the world and NMEA should not be controlling how customers are going to be using it.

Its simple, we provide access to the NMEA 2000 data, we pay NMEA for it. Now NMEA wants our customers to pay them also. Stifling?

Joe Burke
CTO
Chetco Digital Instruments
www.chetcodigital.com

Posted by: oregonjoe at March 6, 2012 12:24 PM

i have been reading over and watching the debate over NMEA2000 since its beginning, mostly here on panbo. the problem with NMEA on the consumer side is the slowness of innovation cuased by NMEA's licensing scheme. i will give NMEA credit. 0183 was a nightmare of cables. NMEA2K systems can be put together and installed safely by the boat owner. however the nmea over wifi/ethernet should make nmea rethink its methods. proprietary systems are simply not doing what consumers need. EVERYONE wants to use a computer/tablet. while i have bought my MFD it does not do everything i need, it was the best at the time, and is my reliable safe navigation tool, i use charts as backup too. NMEA should be making getting into the front door easier for not just the "Industry" players, but also the hobbiest and small innovators. NMEA should not just be working for thier members, standards are to ensure everyone is playing nicely together and protect both the idustry and the consumer. yes i also want certifications, it makes things more compatible, however with as slow as the industry is innovating, as a consumer i no longer care if a product is certified or compliant. i just want it to work, with the features i want. nmea, and marine electronics were very stagnant until recently hit a huge boom, due mainly to moving nmea information into the software/wifi arena. not that the technology wasnt available, just that competition was stagnant. the ipad has made the industry innovate, and NMEA had better take heed. microsoft tried and continues to try to build proprietary standards, and that business model is no longer accepted by consumers. apple and android are going at it, apple has a smoother interface due to certifications after development and before release on the market, however the restrictions and draconian approach to those standards is sending more business to android. right now as a consumer i see NMEA as money hungry, they are attempting to micro manage, thier standard, in the interest of money. thier membership is elitist and mainly composed of industry players, that try to protect thier own interests OVER consumer interests. it appears nmea is the tool to increase profit by stifling competition by keeping out hobbiest and small business competiters. i am currently holding off spending money on new equipment as nothing available now meeds my needs. i want my computer, and chartplotter synched with the same information, i want point and click vhf integration, with plan b redundancy. it needs to be easy to use, i want my wife and kids to easily figure it out, without having to get a industry license. am i asking to much? NMEA you better figure things out!

Posted by: Robert at March 6, 2012 12:47 PM

Joe, It is good to hear you finally at least admit that your gateway products are not simply Actisense NGT-1 technology in new boxes with added WiFi etc. I believe that you've understood how important this distinction is ever since you decided to use only the Actisense chip and not its whole PCB solution. I believe that you knew right then that by the rules you would be building NMEA 2000 hardware products that required a manufacture's code, product codes, hardware testing, certification, etc. Chetco not only knew all that but was reminded of it a long time ago by the Director of NMEA.

But I didn't learn anything about this technical distinction or this history from you, Joe, and that was rather shocking. I came to you as a reporter who was very enthusiastic about your SeaSmart products but concerned about its status with the Standard. I was even sympathetic when you complained about the enormous costs of certification and the unfairness of it since using your products was "exactly like" using an already certified Actisense NGT-1.

As you know, Joe, I encouraged you to bring these matters directly to NMEA's Director and Technical Director, and I even introduced you all via email. It was only then that I learned about all the compliance details and history that you'd failed to mention. I felt conned.

Nonetheless I hoped that Chetco's certification issues would be resolved, and never intended to detail the back story. But after reading all the malarkey you've added to this thread, it's evident that you're still trying play a con game. Your apparent strategy is to strike fear into the software developer community, largely by perpetuating the fiction that your product is just like theirs. Joe, you've known since day one that under the rules your products are defined as "Certified Gateways" and their products as "Approved Software" and there's a enormous difference in how they're treated.

Now you're trying to stir up fear about some "new" restrictions on software developers defined in Appendix H. Well, I'm not sure that Appendix H is new, and I wonder how you would know given that Chetco has only owned the Standard document for a few weeks and that you have spent the last year acting as though even its easily-obtained hardware certification rules were some sort of mystery. Remember, I have the whole email stream between you and NMEA on my desk, and copies of many older emails too.

I don't know what Joe personally said to developers like Jeff Siegel (though the effect was like lighting a stick of dynamite), but in my experience he should not be considered a trusted source. I too would like to better understand the app and software Approval process, and will be happy to report on it, but I don't currently see anything for developers to fear. As stated in the entry, the Approval process is still not finished and Actisense is so far the only Certified Gateway player, but it looks on track and the table above may even somewhat overstate the costs. (Incidentally, Joe, why didn't you mention the low rates Actisense would charge to test your products?)

And there will be other certified gateways. Actually the Maretron IPG100 already exists, though Maretron decided not to put it into the TPG program (and thus are possibly reading this thread with relief), but Navico has announced its GoFree strategy and the latest equipment from Raymarine, Furuno, and Vesper Marine could all be N2K WiFi gateways without even a separate box. A whole lot of NMEA 2000 engineers who are currently working within the NMEA system (but haven't participated in this conversation) are going to be working with apps developers. They already are in many cases.

Let's also acknowledge that NMEA has not hit many companies with a stick. Chetco is an exception, maybe the exception. NMEA has actually tolerated a number of non-compliant small companies over the years and now they're talking about building a NMEA Apps Store once the Approval process is finished. In short, there's no evidence that NMEA will come down heavy on apps developers who don't obtain Approval, and the approach may be more carrot than stick. Approval will be a selling point, especially for software that puts PGNs onto networks as well simply read them.

But lets also acknowledge that up until recently Chetco was not paying a dime to the NMEA, and had not even purchased the minimal documents a software developer like GPSnavX had. However, while iNavX only got a new "N2K supported" feature, additional sales uncertain, Chetco was selling a fairly expensive box that's all about N2K. Remember "New PGN's added weekly"? And while Joe now likes to talk about "working with NMEA" he wasn't even a member last year. He not only failed to contribute anything to the Standard's development, but NMEA oxygen was sucked up trying to get Chetco to play by the same rules other companies do. Just like Panbo oxygen is getting sucked up right now.

Joe, Sorry it had to be this way but perhaps I've clarified my ethical concerns.

Posted by: Ben in reply to oregonjoe at March 6, 2012 1:11 PM

Having sold and installed more NMEA2000 than most dealers, and having been personally involved in correcting numerous HAZARDOUS system designs and installations (by others) I am dumbfounded that safety is not a greater consideration here, if not the ONLY one.

NMEA2000 is a protocol developed to sit on top of a "free" physical layer (after royalties are paid). The specific overarching intent of all of the NMEA2000 restrictions are to ensure that when a certified product is properly implemented, the data gets through EVERY TIME without incident based on a very effective prioritization scheme.

Having read this entire thread along with many others on this subject, safety seems to be the single point that everyone avoids. Or I get to read that it is a red herring from people who have an agenda.

For the record, I have personally seen an errant polling command issued by a software update (to a peripheral connected to a backbone connected unit) max out the network bandwidth. This caused the CANBus based AC/DC switching (with an improperly designed connection to the NMEA2000 backbone) to have to be reset. Yes, the vessel had to go black because of a software glitch from a non-certified device. Although it was improperly connected, the network load was low enough in most cases not to cause an issue. For all you software guys out there 98% compatible IS NOT compatible in the real world.

I have also personally seen a do-it-yourself install with two NMEA2000 antennas tied to a properly terminated backbone, but both of them had the internal terminators activated. The result of this was to damage to CAN port on both antennas and the ONLY plotter on the boat. All needed to be replaced, but for the life of me I cannot figure out how that connection caused the damage that it did.

While most reading this are involved in the recreational end of things, that is not the only master which NMEA2000 needs to serve. When the USCG and IMO look to endorse/adopt a standard, they are unlikely to accept "it works most of the time" or "it shouldn't be a problem". NMEA2000 as designed and certified can meet ANY scrutiny a regulating agency throws at it. THIS consideration should be the only one that determines who is allowed to connect to a backbone.

My soapbox is very old and tired at this point...

Posted by: Yet Another Industry Insider at March 6, 2012 2:12 PM

Gee Ben –

We are a NMEA member.

We have paid all the requested “license” fees.

We have products ready for “certification”.

We are now only waiting for NMEA “volunteers” to complete the paper work.

The recently approved Appendix H description can be publically found at
http://www.nmea.org/store/index.asp?show=pdet&pid=317&cid=7

If anyone knows of a product that currently meets its requirements, I would be pleased to know. If no one else has an approved “TPG” then there can be no “NMEA Approved” apps. This is my point.

We will attend the Northwest Regional NMEA meeting in March to put forward our concerns as outlined. As is our right as a NMEA member we have a voice, we hope to take customer/developer opinions and make them heard.

We were not forced to do any of this but felt it was in the best interest of our customers and developers. You did help push this forward and I would have hoped you were happy with the result.

I will leave the rest for everyone else to decide.

Joe Burke
CTO
Chetco Digital Instruments

Posted by: oregonjoe at March 6, 2012 4:16 PM

For the third time in this thread, Joe, I understand that there is only one certified TPG gateway, the Actisense NGT-1, and the process for granting Approval to apps that use it is not yet finished. So there are no Approved apps yet, but neither is NMEA pushing any apps developers for Approval. Nor is there any sign that NMEA will be aggressive about Approvals (remember that Chetco is exceptional). I don't see anything on your link that seems different from what I've been reporting for several years.

As best I can tell you could be selling a certified gateway right now, or even a certified TPG. Despite all your huffing and puffing about costs -- do you really think there's a person on earth foolish enough to send NMEA extra money in order to tell NMEA that they object to membership? -- they are not out of line for most companies that manufacture electronics.

So here's the obvious question: If you have now decided that NMEA certification will be "in the best interest of your customers and developers", what the heck was stopping you a year ago? It sure would have saved a lot of effort by folks who will never see a dime of Chetco revenues.

Posted by: Ben at March 6, 2012 6:51 PM

Ben/Panbo and Joe/Chetco, it's time to let this go and move forward. Please. As the developer of apps that use N2K, I am sure glad there are gateways out there. I am also Panbo is here to make users aware of them.

Posted by: GPSNavX at March 6, 2012 7:47 PM

"The question is whether or not NMEA is stifling innovation related to NMEA 2000. I don't see it, but I will listen carefully to any and all comments."...
Is the original question posed by Ben!
It appears that some of the comments are starting to get personal without alluding to an answer to the original question.
I would ask, (but I'm NOT the moderator), that we keep the focus on that, and answers that give reason to a stance taken on the original question.
These "vinegar" responses don't help.

Posted by: Paul at March 6, 2012 8:34 PM

All very interesting stuff, I'll steer away from the US politics of it all, however, just a note about the expenses of start up manufacturing, from a commercial marine point of view, not only would a producer have to pay NMEA charges as outlined, but also get the device through Marine Type Approval and for SOLAS either through DMV or Lloyds certification. The expenses are huge and anything that can be done to reduce this IMHO has got to be a good move and can only encourage innovation.

Steve

Posted by: steverow at March 6, 2012 8:40 PM

Thanks, Steve, but can you put any numbers to those other certifications? My sense is that NMEA is reasonable in comparison to other electronics standards (as opposed to software standards, where free & open source actually makes some sense). I'm sure it would be easier to create new products without any standards or certifications at all, at least in terms of financial outlay, but I don't want to live in that world.

Posted by: Ben in reply to steverow at March 6, 2012 8:56 PM

Ben,

I have been reading this thread not as a developer or a start-up but as a boater interested in where the marine electronics industry is going. I have been in the technology business for over 30 yrs and after seeing the SAP, Oracle, Computer Associates, Microsoft and Apple wars, I do not see that NMEA is advancing technology for the better of the marine industry. The argument could be said that NMEA 2000 is better than 0183 only because there is at least a standardized physical layer. The fact that Airmar PB200 was approved by NMEA and then the terminator was changed after the product was in the field made me wonder what happened there. Then the vendors who produce NMEA 2000 software choose to not support the PGNs output - Maretron did not work with the PB200 for all PGNs until several updates.

I think Kee'comments are most on target and if Ben disagrees with Joe - go get a beer and work it out. I am happy to see Cheto and others to join the space simply because they will push the marketplace, just as the iPhone did to the cell phone market.

As a software developer I would love to create a few applications to monitor and manage my boat but the data on NMEA 2000 is no more open than the data on my Raymarine and Lorrance hardware. I have a Maretron and Actisense USB interface to my N2K network but accessing the data is not possible without reverse engineering it.

I really hope Ethernet gets more traction in the marine electronics because there are many more developers and hardware options in that space.

Posted by: HenryD at March 6, 2012 9:13 PM

Let me see if I understand this correctly.

Shell would like to advertise that their motor oil meets Cummins specs. But Cummins won't tell Shell what their specs are unless Shell pays Cummins mega bucks, and promises not to reveal them to anyone else. Then Shell can't use its own lab to test the quality of its oil, but must submit it to Cummins for testing. More mega bucks. Shell baulks, but decides to advertise that their oil meets the Cummins standards anyway, being careful not to claim that it is "approved by" or "certified by". So they get a "cease and desist" order, which forbids Shell from supplying information to their customers, information that they know to be true and accurate. So their customers are left in the dark guessing that the oil is compatible.

Next Shell is accused of operating illegally or unethically, so decides to give into Cummins demands and pay the exhorbitant fees. Of course these costs are passed on to their customers. Boaters get screwed again.

Of course this scenario is entirely fictional, or is it. Change "Shell" to "Chetgo", and "Cummins" to "NMEA".

As a consumer, I don't give a twit whether a piece of equipment meets a particular standard, or is "certified" by someone or not. The only thing that matters is does the item work as advertised, and that it does not create an unforseen hazard. Any attempt to prevent a manufacturer from dispensing true and accurate information about their product is simply not acceptable.

Posted by: Rick R at March 6, 2012 9:17 PM

Thanks for the sentiment, GPSnavX. I have gotten fairly agitated, I'll admit, but I hope you also notice that it doesn't happen very often (perhaps because I'm 65 ;-). At any rate, I may be as stubborn as Joe about seeing this argument through.

And thus I have a question for you, and other software developers: When Joe was encouraging you to spend your time modifying your software to work with Chetco gateways, did he ever mention that NMEA (and Actisense, I think) had warned him that the gateways were not N2K legal and that there would be consequences? If the answer is no, and you knew that Joe knew that all along (which he did), how does that sit with you?

I'm aware that I also have a part in encouraging developers, too, and I apologize. I'm as keen as anyone on getting N2K boat data integrated with mobile apps. But in the future I'm going to look harder at NMEA compliance and I'm going to be less trusting about what I'm told. Hence the agitation.

Posted by: Ben in reply to GPSNavX at March 6, 2012 9:33 PM

Hi Ben
My turn to chime in. As co-founder of Chetco Digital I can speak from the front lines. A year ago we were still developing the final product. I dont know if you ever started a small business in the worst economic conditions since 1930 but we did and cash was and still is very tight since all capital is reinvested in building inventory, R&D and other related items on the P&L. We are not part of a conglomerate nor do we have access to bank captital or VC funding, when we go to a trade show we pay for it out of our pockets. I do know that everything Joe has done to this point has been done via reverse engineering and his extensive background in hardware and software developement.
NMEA related sales were really not a factor in Chetco markets so to dig up 6-$10K to pay for a NMEA certification was not feasible, smart or even possible.
We sell our analog to digital interface to mainly non NMEA related markets such as marine retrofits where no NMEA backbone exists, automotive, oil and gas, mining, remote generators and heavy equipment. So to layout this kind of capital would have been a pretty poor use of company funds.

It was only after I found your blog and Joe started interfacing with the other bloggers that we pursued more aggressively the NMEA product interface. Prior to this we purchased Actisense NGT units as is sold out of the factory as the red dongle.
Since we were selling a product (Analog to digital interface) that was being used by yacht owners in some cases to populate their Chartplotters with analog data we had to use the NMEA2000 "Label" in our marketing to tell people we could interface with the NMEA backbone using the NGT-1 although still at the time many or actually most of our clients did not have a NMEA2000 backbone.
I think if you read all of our press releases dating back a couple years we always stated that the Actisense interface was needed or used. Not until the past year did Joe take the next step to design a more user friendly package.
Early in 2011 we asked NMEA if we could make payments while we worked through the developement.In fact I was teh one that asked if we could make a down payment then make a payment every month. They said NO. We had to pay all upfront. We then asked if we could use the Actisense NGT-1 unit as it comes from the factory but remove it from the red dongle case and insert their PCB into our Remote interface box. NMEA said NO, it had to be external and it had to show externally that it was the actisense device.

So as much as you think you were privey to all our communications with NMEA and others involved you are really only privey to what others felt open to provide you to support their case.What I wrote above may not be what you understood and you may consider me a conman also and thats your opinion and your welcome to it. As we all know what opinions areand what opinions are like.

Joe and I do resent the fact that we (Chetco Digital) were referred to as criminals and unethical by some parties involved. This was uncalled for and immature to say the least. But we are over it and we have moved beyond these rash, unfounded and harsh chartacterizations. We are moving forward with the advancements of this technology and have wasted enough time and brain power discussing it.

I applaude your blog but deplore the fact that your blog is not truely open since you edit entries to fit the day, but it is your blog and you are welcome to do as you see fit and you do. Thank you for your efforts in getting the open standards interface to the forefront and we are confident that by working together in harmony we, you, NMEA, Actisense, developers and everyone else can put the customers best interest ahead of our ego's and attitudes in our efforts to bring cost effective technology to the boating world. It is the customers we must keep at the core of all we develope and distribute.

Thanks again for you efforts.
I think we have beat this dead horse until it just wont bleed anymore.

Lets bring the next round of technology to the masses.
Oh and lastly. Joe and I havent seen a dime of Chetco Revenue's either. It is all pumped back into R&D,inventory, Rent, electric,marketing(trade shows) fulfilling the goverment reporting regulations, taxes, other simple overhead expenses. At the end of the there is no revenue to split up. We could increase the prices I suppose but that would be counter productive and not in keeping with our philosophy. Now that our volumes have picked up considerably with the iPad interest booming, the NMEA 2000 issue basically behind us life is good and customers can count on us to champion their technology needs and expectations.

Posted by: Steve James at March 6, 2012 9:50 PM

Sorry, Rick, but I'm not sure your analogy could be any wider of the mark. But it is a good opportunity to discuss what NMEA is really like, as I've lost track of how many misrepresentations of it are on this thread.

Perhaps the most important thing to understand is that NMEA is not monolithic. The major membership groups are manufacturers and technical dealers. They have some interests in common but there's also plenty of tension between the groups. Anyone who presumes that NMEA works to some grand hidden agenda like keeping "competitors" out (see way above) doesn't know their ass from their elbow.

Even NMEA's manufacturer group is not monolithic, ranging from very small to quite large. But note that even the large companies, like Garmin, are no where near the size of tech giants like Intel, AT&T, and Apple. Which is relevant because in many cases it was those giants that underwrote the "free and open source" software that some folks now presume should be the model for all industry segments.

At any rate, I challenge anyone to look through the NMEA membership -- easily done right here: http://goo.gl/rOnMi -- and to then characterize the organization as a bunch of cigar-chomping monopolists. Look also for Chetco-size companies that build certified N2K equipment without making a huge fuss about it, like Offshore Systems. It's damn hard for any small company to manufacture electronics for any niche, but it is possible.

Posted by: Ben in reply to Rick R at March 6, 2012 10:36 PM

Thanks, Steve. While I definitely do moderate comments on Panbo, I'm pretty sure that's at least partially why they're well read. And whereas I just checked, I can tell you that every comment to this thread is up as written.

But admittedly there's were a few comments by unregistered users that I didn't exactly approve as quickly as possible, like this one:

"Bell Labs had nothing to do with Linux, but everything to do with UNIX."

It takes some digging to understand what that means in this context, or the larger one, but let's just note that there are zealots among us.

Posted by: Ben in reply to Steve James at March 6, 2012 10:53 PM

WOW, this is a dramatic and heated thread, its been going on all day. the question has become lost in the mix. that there are that many unresolved issues to bring this much debate and vinegar, should show something. there is a lot of bitterness toward NMEA. this additude should be a warning sign and partly answers the question. the prices for certification are not an issue. i want protection that the end product works as described, ultimately safety is a concern. on the front end entrance into NMEA hardware and software/app design needs to be as easy and open as possible. on this NMEA IS stifling innovation and competition. the end product is what matters. membership should also not be so prohibitive, separate cost structure for membership would ensure more voices are involved and ultimately better participation and increased innovation. in the end the shift to WiFi and Ethernet gateways is both a sign of the times and an attempt to make things more open. it should be noted that this means that either NMEA is not keeping up with demand, and the new focus in the industry is and will pass up rigid standards. if those standards can not keep up, or are inflexible this has happened before in different industries. this could be simple fixes.

So yes NMEA price structure IS stifling innovation. as a point to consumer demand i bought a Navico 12" MFD as my primary hub for my network, at the time (and still is not) a certified product, its only compatible due to daisy chains. ultimately the features, and flexibility convinced me to spend my money. in the near future IApps and android will be running this show and debate and the innovation they will provide, needs to be made easy and safe or we will be fighting over the next standard.

Posted by: Robert at March 6, 2012 11:03 PM

Ben, NMEA certification or official testing never came up in the discussions I had with Chetco. Instead the exchanges were all technical (that is the side of this I prefer) in regards to working with the SeaSmart API. More nuts and bolts and bits if you will. Chetco was very supportive and even made API enhancements at my request. I learned about the Chetco SeaSmart on Panbo. A good thing. I believe I then contacted Chetco with an interest in adding N2K support to iNavX. Since I already had added N2K support to MacENC and had the official NMEA N2K PGN specification I saw it as a relatively easy task. It turned out to take a bit longer because iNavX supports far more N2K PGNs then MacENC does.

I see no benefit in either you or Chetco "winning the argument" here. Instead let's move forward to the gain of the end user with more connectivity solutions. Of course N2K is not perfect, but it sure makes it much easier to avoid vendor lock which for me is a big deal.

For me it was exciting to see N2K data on with iNavX on my iPhone and iPad wirelessly. Chetco did a nice job with the SeaSmart API.

I have no bitterness towards all involved parties and look forward to working with NMEA, Panbo, and Chetco.

Posted by: GPSNavX at March 7, 2012 12:18 AM

First, i have the utmost respect for Ben and panbo, it has become part of my morning routine. i apologize there is separate pricing schemes, $175 for an associate membership is not horrible. I did go thru the list and Ben is an associate. i can also easily see those names being very much like herding cats. i only found ocean navigator and yachting for media, the majority is for profit, and that is their main objective, followed up in today's economy by not getting sued. i did not see boat us, a large media presence, or consumer entities. it is simple economics that companies would vote in favor of their best interest the majority of the time, i doubt there is a conscious effort, but its too the same effect. $600-$900 for the documentation is a large cost if you wanted to see if an option or idea is viable, and i doubt any of the nmea members would want a hardware geek, designing a cheap device such as a NMEA2000 refrigeration connection, put on the internet, that would cut into profit. and unfortunately yes that does cause a desire to keep the entrance bar up a bit, again doubtful there is a collective thought on this, just simple economics each voting for its own interest. CHETCO is also for profit by the way, and the story does sound a bit fishy, but just a few years ago i don't remember furuno, or navico getting a enforced into the standard. and how many different proprietary NMEA2000 connectors are there now? how many of their products are compliant vice certified?
ultimately the consumer will decide or immediately jump on the next "android" standard that comes along. that this thread has lasted this long and is this heated, should point to the fact that if there was another viable ship a large exodus would take place. and that should be another sign that nmea is stifling innovation. definitely getting off the box now, sorry Ben.

Posted by: Robert at March 7, 2012 12:21 AM

Ben, in answer to your question, MTA for three related similar items £7125 plus VAT. I forgot CE Certification which cost just over £2000+VAT just to get the advice and go-ahead that we could self certify, we have not yet crossed the LLoyds and DMV bridge yet, but as one of our Survey vessels we have an install on, has just got DMV status that is no doubt coming, and if your gear isn't on the list it can't be on there. Bear in mind we are in a very small sector...Survey and Renewables with a small amount of pilotage, and that our top product costs less than £300 without accessories, then you will understand that we have a way to go yet to even think of being profitable. At some future point we will probably be releasing products that do use NMEA 2K so we will add that to the list. Then if we want to sell in the US we will have to appoint a distributor who we will have to pay to get us FCC approval as well, although that will only be required for the Leisure market, as thankfully US/UK have a reciprocal arrangement on MTA. It was never envisaged that we would go into the Leisure side, but that is exactly what we are going to have to do to be profitable (hopefully). One thing that really wicks me off is that If we had been manufacturing in Shenzen rather than keeping the manufacturing here in the UK in Luton, then we could have just walked down the street a little way, and picked up a CE certification for about 100 dollars and a couple of packets of cigarettes, and it would have been acceptable. I notice from the link you posted that there are only eight companies in the UK that are registered with NMEA, how is that so,when the UK is the second largest designer/ producer of Marine Gear outside the US, doesn't stack up somehow.
We have also had enquiries from the offshore Gas sector, but that in itself is a whole new ball game, having to have the gear certified as intrinsically safe in an explosive environment, and probabaly not worth the effort in terms of sales against aggravation.
Sorry if this wanders off thread a bit, but it's important that people know that NMEA approval is not the only setup cost involved in marine gear, and as I said before, anything that can be done to help small startup companies like ourselves and presumably Chetco has got to be welcome.

Steve.

Posted by: steverow at March 7, 2012 5:47 AM

I think there is a fundamental issue here.

NMEA clearly made a decision way back that 2k was to be a revenue raiser. This drives all their thinking. It's primarily why they have got so exercised over gateways as they suddenly saw 2k being bypassed by wifi/tcpip. ( think if you just put the gateway into a depth sensor, poof goes 2k) this is what is freaking NMEA. It's why Appendix H exists.

Look at the huge Open source space both hardware and software it's full of successful companies and multi millionaires. Big companies hated open source initially AT &T tried to close Unix, what happened Linux , where's Unix now, nowhere.

What NMEA missed,rather like Microsoft, was the rapid adoption of wifi/ Ethernet on boats. Then they missed the gateway loophole. Now they are scrabbling.

For example in the telecoms world I can bring a fully certified product to worldwide markets by buying a gsm module fully certified. It's designed to integrate into my pub. If NMEA were serious a simple nod to Actisrnse would have seen this happen.

the fact is NMEA is both Poacher and gamekeeper. It can see its cherished system slip beneath its fingers with wifi and Tcp/ip and it's worried

Today with open source development tools GCc cross compilers manufacturers free design tools, the actual cost to get a prototype to function is way way lower then in the past, witness the myriad of micro companies in the it space. Even CE is free as you can self certify. ( and most Chinese don't even bother as there is no oversight anyway) Fixed cost certification is a dinosaur in this industry.

NMEA have to decide, is its primarily aim to create a worldwide standard for marine networking or is it a commercial undertaking where revenue is the prime driver.

Let's leave aside all the technical issues of which there are many, let's leave aside certification = safety ( it doesn't) Let's leave aside that currently such networks can't scale.

The fact is if Linus Thorvaldsen had decided on "intellectual property rights" where would Linux be now. Or if the Chinese enter the 2k market where IP holds little value. The best defence is to see it succeed hopefully wildly beyond ones dreams, that brings the bucks.

Dave

Posted by: Dave at March 7, 2012 8:07 AM

I'm a senior engineer for a marine electronics manufacturer. I'm sorry, but I must be anonymous.

I live in the middle of this discussion every day. On one hand, I design the hardware and software for NMEA2K devices for a large, worldwide market. On the other, I have my own ideas and plans to create products that my mother company would never create. I live in this small innovation world but understand the standards, big company issues as well.

Anyone reading over this entire thread would walk away with one undeniable fact. Ben was looking for people to defend NMEA as to allowing innovation. This wasn't a quest for information. Ben's continual arguing against a small developer is shocking especially alongside the numerous other developers saying that innovation is hurt by NMEA. Ben, are you listening?

So let me add one more voice. The small volume product ideas I have would be impossible for me to experiment with at home if I didn't have access to all the NMEA info at work. What I'm doing is unethical. If I can build what I think I can, I'll have to join NMEA and pay the full ride to make my product legal. I couldn't do that if I just had a boat and a good idea. Paying hundreds of dollars for documentation in order to allow exploration is not just stifling, it's aggressively anti-innovation.

The small innovator sees needs. We should be encouraged to create, explore, and develop. That isn't being done and it's bad for our industry. What's worse Ben, is that Panbo used to be the place where honest inside exchanges could take place. This isn't happening especially as I hear about postings being removed and edited. It makes me wonder if the picture of the director of NMEA handing you a check lower down on the page has something to do with this. It makes me question the information that is presented on this blog. And after 6 years of reading every posting, that sucks.

Posted by: Anonymous at March 7, 2012 9:22 AM

Panbo is still that place, Anon, which should be pretty obvious way down here in this thread. I'm the same guy, too. Maybe I did get a bit over heated and maybe I even deserve some sliver of your venom. But I'm on a project now and if I do try to untangle your post, it will be much later.

In the meantime, doesn't anyone want to talk about WiFi cameras? http://goo.gl/LDjXq

Posted by: Ben in reply to Anonymous at March 7, 2012 11:12 AM

Interesting commentary from Anonymous et al. I too must remain anonymous because of my involvement with both sides on this conversation as well.

First, the DISGUSTING implication about Ben Ellison's ethics is completely out of line and indicates a sophomoric, uneducated viewpoint. IF you want to hurl stones, know what you are throwing and at whom. The NMEA has paid to sponsor the annual awards breakfast for the Boat Writers Institute annual meeting at the Miami Boat Show each year since 2011. The NMEA began that practice with the specific intent to unveil their Master Dealer Program to the boating public. BWI has a panel of their OWN MEMBERS who vote on the submissions for awards and determine the winners. The BWI Board (I believe) raises award money and allocates it to the winners in each category. To indicate that Ben is being "paid off" by the NMEA is just short of libel, and it may not be short.

Second, I think maybe it is time for all of the independent innovators to come together, coordinate a volunteer effort, develop their own interface protocol, introduce it and market it to the public. Then it would really be free and they would have exactly what they want.

Personally, I am VERY ANGRY that my local Mercedes dealer will not allow me to drive off the showroom floor with a new model when I tell them I just want to use it for a few weeks to achieve my objective (which is to pickup the upgraded supermodel class girlfriend and get invited to swanky parties on Collins Avenue). No matter how much I promise them I will bring it back and they weren't going to sell it for the time I had it out, and I would make sure it would show no wear and tear, they refuse to do so. HOW DARE THEY?!?!?!? Don't they understand that I am just a little guy trying to make my way in this big mean world.

If an innovator has the idea for new products that clients want, find a venture capitalist to sell your plan to. If you need one, let me know, I have access to several and Ben will forward your request to me anonymously.

This is a basic matter of ownership and any other conversation (particularly the insults) is just out of line for so called professionals.

Posted by: Yet Another Industry Insider at March 7, 2012 11:33 AM

Well, there's one thing that deserves clarification right now. To my knowledge there are no comments missing from this thread. None. Zero. And none have been edited.

I did state that I might manage this thread's comments to keep it on topic and in fact I did not publish an early comment by Paul about the New Zealand definition of open standards. You'll see it up there now, right after the one where I asked him specifically not to sidetrack the conversation in that direction.

Let's be clear how all this works. If you're registered on Panbo or have other authentication, your comment goes up immediately. I see it when you see it. If you post anonymously someone has to approve it. Your choice. Obviously, if you go the unauthenticated route, you don't know when your comment will get published. Panbo does not have a 24/7 staff.

In fact I do most of the comment management myself and it's mostly about keeping spam off the site. Sometimes I do actually correct spelling and grammar, mainly to help the commenter get his point across. I really don't like messing with thread content; it's no fun and always annoys people.

At any rate, I gave up on trying to moderate this thread pretty darn quickly. I did let some comment approvals go longer than I had to but it wasn't because they were critical of NMEA or me. It was because they seemed irrelevant to the topic.

That's it, folks, the story of Panbo comment censureship. And, by the way, just look at what's posted up there.

For instance, do I really have to defend myself against the charge that my opinions were bought by a $500 prize I'm proud to have won...because of the random fact that NMEA happens to sponsor the category?

Posted by: Ben at March 7, 2012 12:06 PM

Ben -

I was not aware of the Third Party Certification for apps operating through a certified gateway, but this thread prompted me to read the abstract. I would have read the whole thing but, of course, I'd have to pay for the privilege.

Appendix H does seem to require that any app that uses the information from a NMEA-2000 network must pay up UNLESS the PGN has been stripped out and the message has been translated, not just imbedded, in another protocol. This should be pretty effective at killing off a third party app market if enforced.

Maybe next NMEA could require anybody who owns or is a passenger on a NMEA-2000 equipped vessel to pay a fee unless they are wearing blinders.

A real service from one of the guys on this thread might be to initiate an open standard which could be embedded in a protocol layer over Ethernet, Bluetooth and WiFi. The one "certified" gateway could actually save money because you could avoid the hundreds of dollars in NMEA-2000 cables, connecters and hubs. I'm certain NMEA would try to close any loophole but, as long as the data is translated to a second set of identifiers other than the PGNs, I don't believe a fee could be charged. Perhaps some lawyer here could suggest how much these P*Ns need to be differentiated. Maybe you just need to add a digit to the PGN number.

This of course would all be unnecessary if NMEA where truly interested in promoting the interests of technological advancement, which they are clearly not.

Ben, I know that you've put together a lot of NMEA 2K hardware and I'm really surprised that you haven't run into the electrical and cabling issues of the spec. In even a moderate size cluster of NMEA 2000 certified hardware (like maybe 20 devices) it's so easy to end up with a bus crash. This is a relic of the underlying CANbus. Also, add a couple 20' legs off a hub and put a terminator in the wrong place and God help you. Engine manufactures have taken notice of this. I don't know of one NMEA-2K equipped engine that lets you do more than just observe engine info.

As a developer, I certainly wouldn't want any information that my products used for proper operation to be coexisting on the same wire. The "safety" arguments made by NMEA are self serving and misleading at best.

As for all the "proprietary IP" of NMEA 2000, the standard basically IS CANbus with connectors that are available off-the-shelf. The only thing that's really proprietary is the message content and the PGN numbers. Any old CANbus sniffer can read NMEA 2000. Oops, I think I just tipped off the NMEA police of another criminal.

Then there's the cost of a NMEA 2K network. Instead of buying some RJ connectors and cable at Radio Shack or Home Depot, you probable had to special order your connectors and cables from Hamilton or West and paid $80 for a hub, $18 for a T-connector, $20 to $50 for each device cable, and $10 for each terminator. By comparison, a good quality Ethernet switch should cost $20 or so and each cable should be in the $2 to $4 range.

NMEA should be praised for at least getting the industry together to cooperate on communication between devices. On most other aspects of the standard and for their absurd licensing structure they should be damned for stifling innovation.

Ben, your support of NMEA is admirable but I think you are on the wrong side of this argument.

- Bill


Posted by: Bill Southworth at March 7, 2012 12:10 PM

Thanks "Yet Another Industry Insider", whose comment also illustrates a quirk of the system. He posted it while I was writing my comment and doing other things, so I didn't see or approve it until after I'd finished mine, but yet it's shown earlier in chronological order. That's because comment time stamps tied to when a comment enters the system, not when it's approved, and they are not editable. Unfortunately it can be confusing to readers, and I guess some could see it as evidence of editing that actually isn't happening.

Posted by: Ben at March 7, 2012 12:50 PM

Bill,
I think you have hit the nail on the head there. I dont have much to say about the protocols and layers, because they are essentially ethereal and dont "really exist", or can always be changed. But what on Earth inspired the NMEA peeps to go with an ancient cable technology is beyond me. We got rid of all our resistor terminated coax stuff when we changed from DOS to 3.1.

There are far better cable types out there with multiple cores, cheaply available and would have been when 2K was being developed, HDMI or VGA with triple data screening, being two such examples, and would give longer runs with far fewer packet collisions, than the setup they now have. I'd love to know what drove their thinking on the "physical layer".

Steve

PS Ben this thread is now so long you've run out of "white space" and my PC is now showing all messages with a blue background...lol

Posted by: steverow at March 7, 2012 1:23 PM

Let's call this thread a day and let us small developers go back to developing, even if it is not for NMEA2000.

I'm sure GPSNavX is already waiting for an updated iOS SDK so that he can show charts at 2048x1536 on the new iPad... Overtaking MFDs by miles... Shows the AIS data over NMEA0183 over TCP just fine.

Posted by: Kees at March 7, 2012 2:32 PM


Zealot? I point out an error in one of your comments, and I am a zealot?

Maybe what I should of said was "you might try reading the wikipedia entry for Linux before saying Bell Labs invented it."

Posted by: BrianM in reply to Ben at March 7, 2012 6:08 PM

Actually, Brian, I think I did glance at that Wikipedia entry, or have in the past. I like the family tree of UNIX related software:

http://en.wikipedia.org/wiki/File:Unix_history.svg

Was my sin not noticing that Linux is connected with a dotted line instead of a solid one?

Sorry, I have a hard time seeing how that matters, given that Linux too is underwritten by AT&T size corporations. Have you checked out the sponsors behind the Linux Foundation where Linus Torvalds himself draws a salary?

http://www.linuxfoundation.org/about/members

It seems like open source zealots get jumpy about the corporate underwriting of these standards, but I have no problem with it. I am not a developer but I gather that it's worked very well and I believe that some of the best marine multi-function displays are running Linux. Thank you IBM, HP, INTEL, etc. and the Linux volunteer community.

Unfortunately there does not seem to be any analog to corporate-sponsored open source development in the marine electronics world, and many other small industries. This thread is by no means the first on Panbo where folks have enthused about starting an open source marine protocol, and there's even a long-running thread on the Forum about it: http://goo.gl/sHNS8

But to my knowledge nothing has come of that talk. The exception, I think, is Kees Verruijt, who has apparently deciphered lots of N2K in a legal fashion, mainly so that he could use it to built his own apps. Again I'm not a developer, but I suspect that all the commenters here who are concerned about hobbyist and experimental use of N2K should check out what Kees has made available:

http://yachtelectronics.blogspot.com/

I have long lobbied NMEA to make N2K easier to access for hobbyists and I also have utmost respect for Kees. I think it's quite telling that he's now put his work under a Creative Commons NonCommercial license, so it's freely available for experimenter use but not for commercial products. I wouldn't say that he's enthralled with NMEA but it seems like he's recognizing that their fee model for multi-manufacturer standard is the reality on the ground.

But this thread was never meant to be about hobbyist and experimenter use of N2K. It was meant to be about the realities of company trying to bring electronics hardware into this world. I think that reality is pretty harsh, and that NMEA fees are a relatively minor consideration, but we didn't really get there, did we?

Posted by: Ben in reply to BrianM at March 7, 2012 10:48 PM

Wow this is one Heck of a thread. I will throw this out there from one of my other hobbies. Back in the late eighties early nineties the NMRA (national model railroad assoc) decided to created a net work control standard for model railroads. They openly publish the standard and allow for certification thru minimal fees this idea has allowed for many smaller companies to create products for a relatively small market in this sense it has a lot in common with the marine market. Anyways here is the link to the info, Personally this seems to be what most of the developers on this forum would like to see.
http://www.nmra.org/standards/DCC/DCCHome.html

Posted by: Guest Chris at March 7, 2012 11:21 PM

Very entertaining thread.

Let me add a comment as an occasional customer of boat electronics.

I am delighted with my Ray E120 because it connects seamlessly with my 1991 autopilot. It also talks to a couple of NMEA2000 sensors and hopefully work with any extra sensors I might want to add in 5 years time. I recently added a $1,200 radar dome, and hey, it worked perfectly after I plugged it in.

My 2 month old iPad 2 (which I love) will be obsolete next week. I have no interest in connecting it to my boat electronics nor do I want to set up bluetooth, Ethernet or anything else. This is not an opinion, it's fact. My boat electronics keeps me off the rocks and is not for entertainment.

So for all the comments saying that the NMEA stuff is overpriced, consider what I was able to save by keeping a 20 year old autopilot in operation and then consider all the perfectly good computer equipment you've thrown out since 1991 because it won't play with the latest computer/printer/router/LAN cable/software that you just bought. Innovation does not always work out best for the customer.

Posted by: Pi at March 8, 2012 12:22 AM

Focusing on what’s good or bad with NMEA will surely not solve any of the challenges we are currently facing in the marine industry. It’s a blatant fact; NMEA is outdated and will remain outdated as they have a monopoly in the industry.

Their definitely are major short comings in the N2K protocol but no matter what we are going to have to stick with is for at least a couple of more years as uprooting all networks on board a boat and fitting a new network is surely not going to happen.

With this said, over the past year in collaboration with major players in and outside the marine industry have been working on MARSSA, an open reference architecture and open source platform for marine vessels.

When we say ‘Open’ we really mean OPEN and it’s free for all to use. MARSSA basically abstracts the end developer from the hardware and communication layers, allowing the developer to develop applications which are independent of the architecture on board the vessel by providing an open API and in the future SDK.

MARSSA is integrating with all existent marine protocols BUT we are pushing forward Ethernet and the concepts of a private/public cloud. With such a framework in hand we are enabling an App store concept through which a yacht owners will in the next couple of months be able to install MARSSA on board there vessel and download the necessary apps needed onto their vessel. We therefore encourage anyone interest to join the community and help bring forward this platform.

Posted by: Zak Borg at March 8, 2012 3:12 AM

I've been asking about Appendix H and I'm learning that my TPG reporting to date is pretty much correct. I don't yet know when the Appendix was made "official" but it has been in discussion among NMEA members since 2006. A very reliable source has confirmed that "NMEA will be using the ‘benefits/carrot’ approach to get software developers to become 'NMEA 2000 Approved'."

What I didn't get right is that while the Actisense NGT-1 is a certified gateway it is not yet a certified TPG. That will happen when the Approval testing process is finalized. Then the $99 test certificate comes into play. Actisense plans to charge 120 pounds (about $190) to actually test software using their gateway. Again one of the best places to see the TPG program in progress is Actisense's "compatability" page: http://goo.gl/xiTvJ

The goal of Software Approval is to assure consumers that a given program will not mess with the integrity of their network. I just don't understand how anyone can say that a network distributing critical GPS, depth, AIS, and DSC information is not safety related.

I'd like to add that while I have no idea how Bill Southworth is setting up his N2K networks, the kind of regular failures he describes are NOT normal. I have put together numerous +/-20 node N2K networks, almost always multi-multi-manufacturer and often tested long term in a real boat...and I've seen very few problems, and never a whole system failure. I've also seen how well the cabling holds up. And I've investigated issues like bandwidth clogging ( http://goo.gl/3nYir ) without finding insurmountable problems. I've even invited readers to submit documentaton of real world NMEA 2000 system failures but so far have only heard about ones Yet Another Industry Insider described. True, it's not a good idea to update N2K firmware underway or to put four terminators on a network!

Finally, anyone who thinks that CANbus is some sort of data dinosaur really ought to check out its history and all its "higher layer implementations":

http://en.wikipedia.org/wiki/CAN_bus#Higher_layer_implementations

Posted by: Ben at March 8, 2012 11:24 AM


Hi Ben,

As you know I'm one of the first developers that started working with the Actisense NGT-1, which is the only practical NMEA 2000 gateway as of now.

One of the things that Actisense did early on before the NMEA would assent to the idea of a TPG is that the NGT-1 does not just act as a physical interface translating serial-over-USB into CAN (like most other CAN interfaces). Instead it is 'smart': it enforces that the packets sent by the NGT-1 on behalf of the '3rd party' do not violate the NMEA 2000 standard. In particular it sets rate limits for all PGNs independently. By design it will just not let you flood the bus. It will not let you send "half" packets.

That's all besides it electrical goodness, with it being very careful in that no ground loops/voltage spikes/current loops can occur in the NMEA 2000 network.

The only way you can mess up as a developer is by sending incorrect content, but that's all. But since you cannot override any other sender on the bus, I'd say this is about the only worry a user has.

Let me repeat: there is NO WAY that you can "mess up" a GPS sending on address X or a depth sensor at address Y from the software side of a NGT-1. Your MFD / display device, once set to display data from X or Y, WILL WORK.

Since Actisense did such a good job, why not ease up on the software side?

PS I concur with your findings that, generally, NMEA 2000 networks typically "just work". Some devices themselves may fail, but I've yet to see a network level failure.

Posted by: Kees at March 8, 2012 12:22 PM

Thanks, Kees. My impression is that Actisense designed the NGT-1 in anticipation of the TPG program, but that hardly matters now. I'm really glad to hear that you're so confident about the safety of their gateway, but I can still understand how NMEA would want to see a testing protocol that proves the safety of specific software programs running through it. Then they can confidently issue Approvals that might be reassuring to users, and hence a benefit to the developer. It even makes sense that NMEA standard managers think of such an approval program as an obligation.

But will NMEA force software Approvals? I don't see any evidence of that, and it would seem impossible to justify for read-only apps, but I am asking for clarity. In the meantime let's remember that as best I can tell NMEA has only tried to use force in exceptional situations.

Posted by: Ben at March 8, 2012 1:41 PM

I think you can judge NMEA's success at being innovative with this simple metric:

It is now 2012 and we are arguing about NMEA-2000 as in year 2000. It has taken 12 years for the NMEA-2000 protocol to become sufficiently widely implemented in marine electronics for most recreational boaters to even be aware of it or interested in it.

The notion that the details of NMEA-2000 are secret and only available to fee-paying members is very strange in this day and age. I find that approach to be extremely strange. While I have collected a lot of information about NMEA-2000, it is all from odd channels and random disclosures. Why can't I look up information about NMEA-2000 on the NMEA website?

Do you think we would all be reading this using HTML and HTTPD if the folks at CERN required an investment of $6,000 for each person who wanted to make a web page?

Posted by: Jim Hebert at March 12, 2012 8:49 PM

Jim, I agree that NMEA chose a poor name for the Standard meant to (mostly) supersede NMEA 0183. They probably should named it NMEA21 like the USCG's Rescue21, which was also slow to get started but is starting to shine now. I bet they both last well into the 21st century too.

But I would like to discuss the rest of your comment. From what I've learned over the years the way NMEA 2000 is being managed is only strange if you're unfamiliar with industrial standards as compared to the Internet and computer worlds.

CERN is a great place to start. If you visit its site -- http://goo.gl/k9Uji -- you'll learn that it's an enormous research institution run and mostly financed by 20 nations. There's also some history on the origin of the Web and how it was conceived to share data between CERN and all the universities it's associated with.

So one reasonable question would be what analog to CERN would have or should have developed a core marine data networking standard? I can't think of one with the possible exception of the USCG. But then again I don't think the USCG has anything like the development resources of CERN and, in fact, there's at least one USCG engineer who has been very instrumental in the evolution of N2K.

Another reasonable question is whether anyone would be happy with NMEA 2000 if the standard was as loose as HTML. Heck, we still have to check any changes on this site on as many browsers and devices as possible, and then often adjust. Does that make sense for data critical to navigation?

Posted by: Ben in reply to Jim Hebert at March 13, 2012 12:53 AM

So is NMEA 2000 more like HTML or the SAE1939 standard used in engines (and also built on top of CANbus)? Here's a site where you can supposedly save $1,325 over purchasing SAE1939 documents separately, though they're only for a "single user" and are therefore probably copyrighted. How strange is that?

http://www.sae.org/standardsdev/groundvehicle/j1939.htm

So is NMEA 2000 more like HTML or the DeviceNet standard "used in the automation industry to interconnect control devices for data exchange"(and also built on top of CANbus)? Here's a Wiki description of how DeviceNet is certified:


"1. Register as vendor with ODVA. You will be given a vendor ID.
2. Purchase a copy of the DeviceNet specification. A hard and soft copy will be sent to you.
3. Purchase the conformance test software and corresponding hardware interface card. Note that only selected interface cards from a few vendors can be used.
4. Develop and test product in-house. You would probably need help from the discussion group, see the External links below.
5. Submit your product to ODVA test lab for independent verification.
6. Repeat the above two steps until your product successfully pass the independent test."

Sound familiar?

http://en.wikipedia.org/wiki/DeviceNet

Those are the first two CANbus "higher layer implementations" I looked at, and it seems very likely that others also have costs and certification procedures like NMEA 2000. Except, of course, for SmartCraft, which really is a closed standard.

http://en.wikipedia.org/wiki/Canbus#Higher_layer_implementations

Plus there are many other industrial standards families besides CANbus, but I'm tired...

Posted by: Ben in reply to Jim Hebert at March 13, 2012 1:17 AM

On a historical note, I remember using a previous system called JANET, which Uni's and larger Schools and Colleges had. It was a collection of BBS's and Nodes, housed at various edu establishments throughout the world. It worked well, and allowed the sharing of information and transfer of files and data albeit at speeds not much above 9k6 in the mid-eighties. TBL invented the internet as we know it today with http, but the structure to enable that to happen was already there. A similar structure was also available in the Amatuer Radio Packet community, and had links into and out of JANET. So it was very possible even in the 1980's for enthusiasts to send each other programs and just about every other form of data and email. Most of this was done using a protocol called X25 or AX25, later to be superceded by TCP/IP.

Steve

Posted by: steverow in reply to Ben at March 13, 2012 7:58 AM

My comparison of NMEA-2000 to HTML was to point out the speed at which the two standards spread and were adopted. Had the creators of HTML and HTTPD chosen to pursue an expensive licensing structure, the standard would not have spread and flourished as it did when the creators made the decision to openly disclose the method. HTML went around the planet in about 12-days; NMEA-2000 has taken 12-years and is still in its initial stages of adoption by recreational boaters. Yes, CERN was a sugar-daddy for the initial development of HTML. Maybe NMEA needs a few sugar-daddies from the marine electronic industry.

My comment about "strange" is from the perspective of a consumer. If I buy an outboard engine that says it is NMEA-2000 certified and I network it with a display that is NMEA-2000 certified, I would expect to be able to know a little bit about the data the engine is going to send to the display. For example, I would expect to know a few details like that parameters are sent by the engine and what parameters the display is willing to show. As a consumer, in most cases I cannot locate this information. I am just told the devices are NMEA-2000 devices. In a few cases I might be able to dig up a PGN number from the manufacturer. For example, SIMRAD recently began to include PGN numbers is some of its documentation for its displays, and Evinrude began to mention some PGN numbers in its ICON gauge literature. Those numbers tell me very little, unless more details are given. If I search around the web I might be able to find a brief description of the parameters sent in the PGN on some enthusiast websites. These enthusiasts all have reverse-engineered the protocol--because if they bought the information from NMEA they could not disclose it--a real Catch-22. I believe a lot of this obscurity is an artifact of the NMEA's insistence on non-disclosure and in keeping the protocol details hidden from non-paying members. NMEA ought to publish a comprehensive list of PGN numbers and brief details of the data sent in those PGN's on their website. Manufacturers could simply specify what PGN's their devices work with, and consumers could make informed buying choices.

Posted by: Jim Hebert at March 13, 2012 12:14 PM

Good grief, Jim. NMEA has a searchable database of PGNs I've often referenced:

http://www.nmea.org/content/nmea_standards/messages_pgns.asp

I do wish that public database were more detailed but you can find higher PGN detail in a number of places. One is the manuals of NMEA companies like Maretron and Airmar. Another is by using Maretron N2KAnalyzer or Actisense NMEAreader software with the respective company's gateway, which will let you view most every detail running on your backbone. Similar is true of several MFD's, like the Simrad NSx line, which will also detail incoming PGN fields while showing you the live data simultaneously.

Moreover, product PGN transmit and receive lists are not random. I believe their documentation is required by the NMEA certification process. If your outboard is N2K certified but doesn't specify PGNs send and received in its manual why don't you write NMEA and complain? Just remember that the few "sugar daddies" in marine electronics are already largely financing the organization that will try to answer your question.

Should I presume that you couldn't name an analog to CERN or the university system or the huge tech firms that have all underwritten web and other developments -- which is wonderful, by the way -- because there are none in the little world of marine electronics?

But why no response to my research about how standards similar to N2K are managed? If your boat engine was J1939 or SmartCraft instead, would you expect details about their data outputs for free? Do you think Mercury or Caterpillar or Volvo would let you develop hardware that integrates with their engines without testing? Who would be the sugar daddy in that situation, or should any company developing a critical electronics product to a standard expect compliance to be part of the startup cost?

Besides, isn't the central goal with all these engine standards simply delivering motors, controls, and gauges that work? Isn't that what most consumers are looking for too?

What's different about NMEA 2000 is the promise that products from different manufacturers can work together, including small manufacturers. I see that in action every day I'm on my boat. But who said that NMEA should or could somehow drop the copyright and certification management style used in similar standards situations? Boat data networking is so not HTML.

Posted by: Ben at March 13, 2012 1:45 PM

Again you beat me to an answer Ben! ;)

Although I was going to reference the NMEA's simple pdf of the top-level PGN details as I didn't know it had evolved in to that great searchable database option.

The NMEA should publise that link far more.

Posted by: Andy Campbell at March 13, 2012 1:52 PM

Good grief, Ben, that NMEA data base you point to contains very little data, and you can't tell me that it's been available since 2000. It has to be a quite recent addition.

Posted by: Anonymous in reply to Jim Hebert at March 13, 2012 11:21 PM

Ben--How would I possibly be able to know that it is a requirement of the NMEA-2000 standard that a manufacturer tell me about the PGN's their device can use? I have not purchased a copy of the standard. In order for me to know that a NMEA-2000 certified device must inform me about its PGN's, I think I would have to purchase the standard to know that, wouldn't I?

And Ben, it is not proper for you to tell me details of the standard like you have done, if you have purchased the standard yourself, because you are not supposed to disclose the details. Isn't that right?

Do you see how strange this non-disclosure requirement becomes?

Posted by: Anonymous at March 13, 2012 11:27 PM

"How would I possibly be able to know that it is a requirement of the NMEA-2000 standard...?"

Umm, by reading public documents like the "NMEA 2000 Certification Process" I already linked to in this very thread: http://goo.gl/fzAqh

Page 5:
"Once certified the manufacturer:
• Shall publish the transmit and receive list of its respective product’s PGNs. This could be on the website or in owner’s manuals. At a minimum the manufacturer shall list the PGN numbers. If this list is in electronic format, the manufacturer may link to the NMEA website for descriptions."

I don't own the NMEA 2000 Standard; I just did research instead of presuming negatives. Have you even looked at the N2K White Papers, Anon?

http://www.nmea.org/content/nmea_standards/white_papers.asp

Also, the first Panbo link to NMEA's online PGN database on Panbo was in a June, 2009, entry about an actual N2K problem:

http://www.panbo.com/archives/2009/06/nmea_2000_ais_not_yet_right.html

Posted by: Ben in reply to Anonymous at March 14, 2012 6:54 AM

Seems like the J1939 standard is a lot less expensive then the NEMA standard.

http://store.sae.org/jpaks/

indvidual sections can be purchased for $50 and $615 buys enough downloads to at least create a product.

Also you failed to mention MODbus which is the most used CAN standard behind CANopen (j1939 is part of CANopen)which is a fee and open standard. from their site

http://www.modbus.org/faq.php

Open: The Modbus protocol was transferred from Schneider Electric to the Modbus Organization in April 2004, signaling a commitment to openness. The specification is available free of charge for download, and there are no subsequent licensing fees required for using Modbus or Modbus TCP/IP protocols.

MOD bus also has an ethernet version that works well personaly I think maybe some of the developers on here should get together and create a NEMA alternative using something like MODbus thus creating a more open standard to move forward with.

Posted by: guest chris at March 14, 2012 11:57 AM

Gosh, Chris, I did mention that I had only researched "the first two CANbus higher layer implementations" and I provided a direct link to the Wiki list.

What you failed to mention is that $615 only buys 10 downloads from a library of "SAE's 2,400+ ground vehicle standards". How did you know that 10 are enough to create a product? If you have actual experience with that, can you please also tell us what else it costs to get such a product into SAE compliance?

At any rate, thanks for discussing some standards that are more relevant to NMEA 2000.

Posted by: Ben in reply to guest chris at March 14, 2012 1:31 PM

I didn't realize until today that the PDF version of NMEA's PGN list includes field definitions. It's easily downloaded here:

http://www.nmea.org/content/nmea_standards/downloads.asp

Posted by: Ben in reply to Jim Hebert at March 14, 2012 1:33 PM

Didn't mean to beat you up Ben. I didn't realize there was quite so many parts to j1939 I just did a quick look at the pages and figured with the base standard the physical standard and a few others that apply you would get what you need. Having not read the standards of course this is a complete assumption on my part.

Posted by: guest chris at March 14, 2012 1:47 PM

No wonder I couldn't find much on J1939 when I was curious last year.

Posted by: Dan Corcoran (b393capt) at March 14, 2012 10:20 PM

I think this thread is caught up in the minutia and marketing around N2k. Everyones needs to step back and look at it from a macro point of view.

I personally have been involved in many industry trade groups and standards bodies in various disciplines during my career, and one thing you must realize is that these groups, NMEA included, are not some God that is all knowing and all seeing. They are, in fact, a bunch of purveyors of goods who by participating in a collaborative organization believe they can increase their own revenue. The NMEA membership controls the work performed and the output. They do not participate out of a desire to help mankind. The blah, blah about safety, etc. is simply just marketing.

NMEA constructed the N2k spec based on their perception of consumer demand. But their underlying motive of revenue/profit was by far the biggest driver in the architecture.

I’ll forego sighting obvious N2k attributes that convince me that ‘keep marine electronics proprietary’ was an overarching requirement.

Open systems scare the NMEA membership, they do not want to face the commoditization of their product lines that true open systems would bring. The end result architecture tells the tale. N2k is a ‘controlled’ specification, not an open specification.

Posted by: DotDun at March 15, 2012 12:07 PM

I hardly dare comment on this thread. Innovation equals change. Sometimes stability is more desirable. NMEA 2000 is hardly established enough to be overly concerned with change. Having said that, their slowness has meant that obsolescence is a constant threat. Things evolve. Some old species continue to thrive along with the new, others die out. Which brings me to the history of computer products which are mentioned as examples. Nobody remembers what Ethernet was like before, or remembers its competitors. Nobody remembers what the products were which HTML replaced. Nobody remembers failed standards or attempts to create standards. For all the successes in the computer world, there have been many more failures and many successes which did not survive long. There is also the other extreme -- things which fork or branch out into "similar but different". Unix is not just one thing. Neither is Linux. It's hard to be compatible with everything in the long run. That's why standard writing is hard. NMEA might have been too naive going into writing NMEA 2000. Oh well. I'm for making the best of the situation, however imperfect it is. It still beats each manufacturer having its own standard (with a new one every few years).

Posted by: norse at March 15, 2012 1:52 PM

Very well put Norse. Sums up the situation precisely.
Although I do remember the competing systems in the 70's and 80's, only too well unfortunately and the nightmare of getting them to talk to each other.
Although I have forgotten how to make an 8/16bit parallel system talk to RS485, but I used to know how,.. once upon a time.

(is this ever going to end?...perhaps not).

Steve

Posted by: steverow in reply to norse at March 15, 2012 3:22 PM

I'm late to the game (I missed out on all the "live" fun above), and slightly off topic, but I think NMEA 2000 is a significant advance on what we had before. It has enabled me to implement an information and control system on my yacht that I would have only dreamed about 5 years ago, and it's very robust and I use components from different suppliers. If I had written that sentence 10 years ago, you all would have said that I was dreaming.

Does it stifle innovation? Yes and no.

Clearly it has resulted in significant forward progress, which would say that it has helped innovation.

Have some developers not been able to participate? By the sounds of it, yes, and in that respect it has stifled some innovation. But I don't see that as necessarily bad.

I own a small software company. Laissez faire open systems can (and do) create significant forward progress, but in my opinion are often not very end-user (read "consumer", not "tech-geek") friendly. It takes effort, skill, and money to create good end-user products. I don't see a $6,000 entry fee as a barrier to entry that is significant enough to deter good progress.

And NMEA 2000 will evolve, and yes, may be replaced at some point with something better. That has always happened, even to the best "standards", whether they be open or not. That will have taken some people effort, skill, and money, and thus I will not begrudge their wanting to get a return from that.

The leisure marine industry is a very fragile and niche animal. We have seen that countless times with reputable and significant players going under (we've seen that very recently with Raymarine).

Despite this, we have seen some really significant advances in MFDs and other electronics. Is it as bleeding edge as it could be? Possibly not. But I, for one, am very pleased as a consumer with the advances.

NMEA 2000 has very positively affected this environment and I am very happy that a small group of people took it upon themselves to bother to do it (I know how hard it is to develop industry standards: I worked on 2 standards in mobile computing through the 90's).

It's very easy to throw stones at a system that already exists (none are perfect, except those that we develop ourselves, of course). But I doubt that many of the stone throwers could have developed something that has advanced my electronics ( as a consumer) as much - oh, hold on, none of them actually did. But it seems like some want to ride on the coat-tails: isn't that wanting your cake and eating it, too (excuse the mixed metaphors).

Meanwhile, I say thanks to those who persevered and created forward progress which has given me some really cool new stuff on my yacht. :)

Posted by: Taniwha at March 20, 2012 2:20 AM

I’m late, but this is such high quality discussion and I wanted to share my view as well.

Ben’s question: Is NMEA stifling innovation?
My answer: Yes, in many ways, but the entry fee is not the most important one.

N2K is lacking market interest. E.g. look at Google Insights for NMEA 2000 and NMEA 0183 [1] . There is a constant declining interest in NMEA 0183 over the last six years, but very little increase in NMEA 2000. The market is not adopting NMEA 2000, leaving a big vacuum. From Google Insights I also conclude the only country interested in NMEA 200 is the US. As the Marine tech industry is already a niche market, NMEA has been able to position N2K in the niche of the niche. NMEA (and it’s members) have some work to do. Also compare NMEA v.s. AIS. We all understand the uptake of AIS over the last years, but where is NMEA (2000)? Is AIS that more important than NMEA (2000)?

The success of open standards (and open source as well) can be measured by adoption and size of the community (in the widest sense ranging from users to contributors). How well is N2k adopted and how big it the community? David Wheeler wrote: Successful open standards have FLOSS implementations [2]. An open standard and a FLOSS implementation would allow an active debate by every expert in the field to drive the standard forward. This increases speed of adoption, faster evolution of the standard and faster and increased growth of N2K experts (the N2K community). Wide adoption and large community will not only improve N2K itself but also increase innovation of the products relying on it. However N2K is hardly adapted by the market and that’s a serious issue.

N2K has been around for many years and as we all can read is not mature. Looking at the dates for the N2K Standards Corrigenda [3], progress and evolution of N2K is extremely slow (like its adoption). The closed nature and very small size of the community (i.e. NMEA members) is debit to this. Why are the paying members accepting this immature standard and slow evolution, while on the other hand there is a huge vacuum created by NMEA? Coming from a 100% proprietary closed/lock in model this might be a big and scary step for them, but they have to wake up. Consumarization [4] is everywhere, also in the marine world and adoption and innovation move at an astonishing rate. Look at the success of tablets and social media for professional use. Also crowdsourcing [5] is used by innovative companies (Activecaptain being a good example [6] ) to build a contributing, connected community that adds value and everyone is benefitting from it. If the NMEA and its members are not going to move, I predict a difficult future for them.

An earning model based on paying for specifications of the standard which are distributed via PDF files makes me sad. However its fully in line how the NMEA operates on other areas of N2K promotion, evolution and adoption. Have they ever seriously considered other earning models? How have iTunes and Spotify changed the music industry? What are you paying for the services delivered by Facebook or Google? How is work on Linux and Firefox funded?

NMEA has setup the N2K organization as a cathedral while the marked is clearly ready (I should rather say hungry) for a bazaar approach [7]. Are marine electronics manufacturers really satisfied with this slow adoption and evolution of a immature standard that is over a decade old? They are investing millions to get their products N2K ready and certified. This could be seen as the N2K/NMEA lock in. They will stick with it no matter what. Isn’t there a large group of manufacturers that could build more innovative products that boaters relay want to spend money on, if the standard was more mature and widely adopted (and not only by the marine dinosaurs).

NMEA should make N2K a hot standard everyone wants to have and be in compliance to. Not a standard that is rebranded by manufacturers. The entrée fee is by far not the biggest problem that is stifling innovation. It’s the structure and decades old processes of the NMEA that is holding back N2K progress and a large part of the marine tech industry. Imagine how live would be if the standard was as mature and widely adopted as open standards used by consumers? Not possible? Step back and think again. How can we make this possible? Maybe we can start a topic with advise to the NMEA.

NMEA and members, wake up! There is some work to do. N2K is a very good initiative but you guys need to speed up (a lot). If the vacuum grows we all get sucked in and N2K is dead(, probably eaten by a real open standard with a FLOSS implementation.)

[1] http://www.google.com/insights/search/?hl=en-GB#q=nmea%202000%2Cnmea%200183%2Ccanbus&cmpt=q

[2] http://www.dwheeler.com/essays/open-standards-open-source.html

[3] http://standards.nmea.org/NSNA/corrigenda/nmea-2000/

[4] http://en.wikipedia.org/wiki/Consumerization

[5] http://en.wikipedia.org/wiki/Crowdsourcing

[6] http://www.crowdsourcing.org/site/activecaptain/wwwactivecaptaincom/5631

[7] http://www.catb.org/~esr/writings/homesteading/cathedral-bazaar/index.html

I’m working for fast growing commercial open source company and hold university degrees in electrical engineering and computer science and I’m specialized in open standards and protocols.

Posted by: Paul Jongen at March 20, 2012 5:25 PM

Paul, I don't know if you get around boats much but that would be a better way to gauge the success of NMEA 2000 than Google Insights. Please remember that most boaters don't give a hoot how their electronics work as long as they work. Some do value a multi-manufacturer sensor data standard and thus N2K support is featured to some degree but at this point it's pretty much expected. There are some exceptions like the majority of fixed VHF radios, but I'll bet they adopt N2K over time.

I also suggest that you look at some of the industrial standards mentioned here, which are certainly open to use by any manufacturers but are certainly not FLOSS. They're doing fine too. They are what NMEA 2000 is modeled on and are neither cathedral nor bazaar in structure.

I wish that NMEA 2000 standards making would progress faster, but I can't see how that's helped by companies that try to avoid supporting the standards making organization, which, as you note, is not very expensive to support.

Posted by: Ben in reply to Paul Jongen at March 20, 2012 6:12 PM

I hope some folks on this thread will take a look at Teleflex Marine Optimus 360. It's a good example of why NMEA chose CANbus to build NMEA 2000 on top of. Also a good example of how the two can work together.

360 is a very innovative engine control system that will work with a variety of autopilots and other electronics right out of the box, with one N2K/CANbus cable. Thus consumers will see competition among the peripherals right away, despite all that nasty NMEA monopoly BS you read above. Sorry, no FLOSS or HTML in sight either ;-)

http://goo.gl/uSuHt

Posted by: Ben at March 21, 2012 4:50 PM

Chris, I looked into MODbus as you suggested and it doesn't seem to be quite as free and open as you might think. General Membership for a company under 100 employees is $1,500 though that does include the TCP toolkit, which is $500 otherwise. And MODbus conformance testing runs from $1,350 to $1,600 per device.

Most interesting is the EULA you have to agree to before downloading the standard. Schneider Automation still claims copyright to it! http://www.modbus.org/eula.php

Also do you realize that Schneider built and used the standard for 25 years before turning it "opening" it up to the non profit? Or that Schneider's current 20 billion euros in gross sales is about 26 times more than the entire global recreational marine electronics gross?

Also, did you see what CANopen development and testing software goes for? Makes NMEA 2000 stuff look kind of trivial: http://goo.gl/La79q

Posted by: Ben in reply to guest chris at March 21, 2012 7:33 PM

Ben,

Are you sure the Teleflex 360 canbus and the N2k will be connected on the same cable?

Not much info available on the Garmin or Raymarine web sites yet, but I find no mention that both networks will exist on the same wire. Also, I would imagine NMEA might have an issue with that.

We'll have to wait until the APs ship.

Posted by: DotDun at March 21, 2012 9:43 PM

Did you read the Optimus 360 entry, DotDun? I didn't say that CANbus and N2K would be on the same cable; what I wrote was this:

"And speaking of innovation I couldn't help but notice how well NMEA 2000 works to help make a radically new technology like this possible. I'm not sure where the CANbus networks in Optimus 360 end and the N2K interface begins but since N2K is just a layer on top of CANbus it wasn't hard to do. Isn't that why Garmin has already introduced a GHP20 Steer by Wire autopilot that works with 360? Raymarine has also announced the SPX40 CANbus AP and I heard that Simrad and others are on the way."

I don't know if the AP connects to 360 with a CANbus cable or a NMEA 2000 cable, though of course they can look exactly the same because the DeviceNet cable standard is used for both. That's why I referred to "one N2K/CANbus cable" above.

Posted by: BenTest in reply to DotDun at March 21, 2012 10:17 PM

Ben,

I read what you wrote, then read the features of the 360 system. With the limited amount of technical detail released, an early conclusion is that 360 cannot run on NMEA 2000.

If they truly are separate networks, how did NMEA 2000 "help make a radically new technology like this possible" ?

I don't understand the advantage of using the same 'type' of cable to make disparate connections. In fact, I would suggest that to be a disadvantage as there could be errors in connecting devices to the wrong network.

Of course, if N2k had been implemented on IP, then disparate systems could share the network layer. But, you barred having that discussion when you started this blog entry.

Posted by: DotDun at March 21, 2012 11:32 PM

You asked:

    "The N2K WiFi gateway issue, is NMEA stifling innovation?"

The absurdly obvious answer is a resounding "YES!"

But of more concern to me right now is the "why" behind that answer, and especially your rather transparent attempt to "stack the deck" of the ensuing discussion:

    "I also don't want to see the discussion sidetracked by the argument that NMEA is not an open standard because it's not free. The base definition of an open standard is one that doesn't exclude any manufacturer -- like N2K -- and while "free" is common in the computer world, fees are common in the industrial world that N2K more closely resembles."

Your definition quoted above is perversely twisted, and that skews this entire discussion. Therefore, this issue is NOT a "sidetrack"; it is the very crux of the biscuit, upon which any *meaningful* further discussion must be based, or we're just going around in circles (and based on having read the entire thread-to-date, I'd say we're on about the 180th lap). So, my apologies in advance for declining to forgo this aspect of the debate; but I feel it is too essential to ignore. It is partially for this reason that I am submitting this response before I officially "join" Panbo, so you will have an opportunity to read and consider it before it appears publicly. I do hope you will choose to publish it; but that, of course, is up to you. So, with that out of the way...

To call NMEA 2000 an "open standard" is preposterous on its face. It is barely a "standard", and is "non-open" in the extreme. The very fact that I, as a private individual boat owner, cannot even read a copy of the standard (at least without prohibitively exorbitant cost, which effectively means not at all) tells us all we really need to know about that. But even beyond that, the NDA requirement is incontrovertible proof of the "closed" nature of the standard. Such restrictions are simply not compatible with ANY "open" model, period; and more to the point, they inherently form an (obviously deliberate) barrier to entry. Reference any IEEE or IETF standard for a glaring contrast to the tightly locked-down NMEA model. Can you say, "Good Ol' Boys Club"?

As for the term "free", and its specific relevance to this issue, you apparently need to bone up on the difference between "free, as in beer" vs. "free, as in libre"; try this: http://en.wikipedia.org/wiki/Gratis_versus_libre . All the subsequent dickering about whether NMEA certification costs are or are not "too high" (whatever that might mean in any specific context, which is enough of an intangible moving target to be a perpetually unanswerable red herring in the general case) completely misses this point. Testing and certification fees are a separate matter from whether or not the standard itself is or is not "open". (And for the record, I think it is perfectly reasonable for there to be such fees, perhaps even substantial ones; but that is another discussion entirely.) The standard itself, however, MUST be fully and publicly published (with little or no restriction on its subsequent redistribution) to have any pretense of being "open". The relevant documents don't necessarily need to be "free, as in beer"; but if/when the cost significantly exceeds the direct duplication and distribution costs of the underlying media, it is no longer an incidental "cost covering" fee; it is a (dearly) paid licensing arrangement, by any name -- i.e., the exact antithesis of an "open" standard. If you have ANY doubt about this, go back to that Wikipedia page you later cited: http://en.wikipedia.org/wiki/Open_standard , and re-read the first sentence: "An open standard is a standard that is publicly available ...".

If that still isn't enough to change your mind, keep reading that same page. In particular, look at virtually any of the individual formal definitions cited; you will find statements such as:

  "'Open Standards' are standards made available to the general public ..." (ITU-T)

  "The standard has been published and the standard specification document is available either freely or at a nominal charge. It must be permissible to all to copy, distribute and use it for no fee or at a nominal fee." (EU)

  "An open standard is accessible to everyone free of charge..." (Danish government)

  "...whose specifications are public and without any restriction in their access or implementation." (French law)

  "An open standard fulfills the following conditions:
    * it is public, and its use is available on a free [gratis] basis, or at a cost that does not imply a difficulty for the user.
    * its use is not subject to the payment of any intellectual [copyright] or industrial [patents and trademarks] property right. (Spanish law)

  "Open standards: technical specifications, published and ... available to everybody ..." (Venezuelan law)

  "... a standard shall be considered open if it meets all of these criteria ...
    3. open access: all may access committee documents, drafts and completed standards free of cost or for a negligible fee.
    4. It must be possible for everyone to copy, distribute and use the standard free of cost.
    5. The intellectual rights required to implement the standard (e.g.essential patent claims) are irrevocably available, without any royalties attached.
    6. There are no reservations regarding reuse of the standard."
  (South African Government)

  "* Be accessible to everyone free of charge: no discrimination between users, and no payment or other considerations should be required as a condition to use the standard.
  * Remain accessible to everyone free of charge: owners should renounce their options, if any, to limit access to the standard at a later date.
  * Be documented in all its details: all aspects of the standard should be transparent and documented, and both access to and use of the documentation should be free."
  (New Zealand official interoperability framework definition)

Heck, even Microsoft, one of the most infamously anti-open-standards (and anti-competitive in general) organizations known to mankind and the birthplace of the "Embrace, Extend, Extinguish" mantra, at least gives lip service to this same fundamental concept: "...'open' refers to it being royalty-free, while 'standard' means a technology approved by formalised committees that are open to participation by all interested parties and operate on a consensus basis. An open standard is publicly available, and developed, approved and maintained via a collaborative and consensus driven process."

So how long are you going to cling to the delusion that "all the soldiers are out of step except my little Johnny"?

Another point that seems to be getting missed a lot here is that these devices *are* (at least primarily) computers, not electro-mechanical industrial devices. Hence, that is -- or at least "should be" -- the applicable model, particularly as the underlying platforms inevitably become more and more generic and interoperable with each passing day.

This is no longer 1957. It's time for the NMEA (and you?) to understand that the world *has* changed.

Posted by: NotYetJoined at April 3, 2012 2:47 PM

Funny you didn't quote any of the Wiki definition -- much, much debated I'm sure -- that disagrees with you, like the first paragraph:

"An open standard is a standard that is publicly available and has various rights to use associated with it, and may also have various properties of how it was designed (e.g. open process). There is no single definition and interpretations vary with usage.

http://en.wikipedia.org/wiki/Open_standard

Well actually you did, since NMEA 2000 is both publicly available and royalty free.

My favorite part of your screed, though, is the "these devices *are* (at least primarily) computers, not electro-mechanical industrial devices." That's very funny.

You're confusing the MFDs that often manipulate and display N2K data with all the sensors that deliver the data. I was just looking at diagrams of a big install yesterday, maybe 40-50 devices on the N2K backbone (which has a bridge and multiple power feeds). 90% of them are relatively simple electronic devices, not computers. They are devices typical of the CANbus world.

The devices represent at least three manufacturers and there's lot of data sharing going on, all possible because N2K is open to all manufacturers. There's a lot of innovations in the system too. And it definitely bridges out a PC world of Ethernet and WiFi and apps, just as N2K was conceived to do from the get go.


Posted by: Ben in reply to NotYetJoined at April 3, 2012 4:39 PM

Ben said --

    Funny you didn't quote any of the Wiki definition -- much,
    much debated I'm sure -- that disagrees with you, like the
    first paragraph:

    "An open standard is a standard that is publicly available
    and has various rights to use associated with it, and may
    also have various properties of how it was designed (e.g.
    open process). There is no single definition and
    interpretations vary with usage.

    http://en.wikipedia.org/wiki/Open_standard

I don't see how that disagrees with my point -- unless perhaps you take just the last sentence, out of context, and use it as an excuse to ignore the entire thrust of the article as a whole, including all the formal definitions I cited earlier.

So, while it is true that "[t]here is no single definition and interpretations vary with usage", one thing they ALL have in common is that the standard is openly published and freely available to the general public. That is an essential characteristic of an open standard, and not even close to what NMEA does. The very fact that NMEA considers (or at least claims) the N2K "standard" to be their "intellectual property" puts it in direct and irreconcilable conflict with the concept of an "open standard".

    Well actually you did, since NMEA 2000 is both publicly available
    and royalty free.

Really?!? Where? Can I perhaps go to my local public library and read the full specification there?

According to the NMEA web site http://www.nmea.org/store/index.asp?show=cprd&cid=7 , just to download (no paper, hence no copying costs) the so-called "Full Set" (which really isn't) would set me back at least $4,000 -- onerous (if not outright prohibitive) for most normal mortals -- and even that doesn't get me the details of the test procedures or criteria, to know what the various tolerances really are. For *that* information, I have to sign an NDA. So how exactly is that "publicly available and royalty free", again? Even if we ignore the NDA aspect (how can we?), four grand seems like a heck of a royalty, to me.

    My favorite part of your screed, though, is the "these devices
    *are* (at least primarily) computers, not electro-mechanical
    industrial devices." That's very funny.

How so? I meant exactly what I said; and I wasn't joking.

    You're confusing the MFDs that often manipulate and display N2K
    data with all the sensors that deliver the data.

Not at all. In fact, from the above, it appears that it is you who is confused. Just because a device doesn't have a keyboard or a big fancy LCD display does NOT mean it's not a computer. Read on...

    I was just looking at diagrams of a big install yesterday, maybe
    40-50 devices on the N2K backbone (which has a bridge and
    multiple power feeds). 90% of them are relatively simple
    electronic devices, not computers. They are devices typical of
    the CANbus world.

And each and every one of those devices -- such those Lowrance fuel-flow sensors I mentioned in another post, just to use a handy example (BTW, I typoed the model number in that post; they were really EP-60Rs) -- are in fact computers themselves. They are very small and highly specialized computers to be sure, but computers nonetheless. They include such recognizable subsystems as analog/digital converters, internal memory, a processor of some type (with an instruction set), and a network interface. How is that not a computer? This is further self-evident from the fact that they output *digital* data and "talk" over the N2K bus. Put simply, *ANYTHING* (except the raw power feed, of course) which attaches directly to a CANbus network (including the N2K variant thereof) *IS* a computer, at least in the broad sense of that term (to be somewhat pedantic, a more precise term would be "microcontroller"; but as a practical matter all modern microcontrollers are in fact computers anyway -- cf. http://en.wikipedia.org/wiki/Microcontroller "A microcontroller ... is a small computer on a single integrated circuit containing a processor core, memory, and programmable input/output peripherals").

    The devices represent at least three manufacturers and there's
    lot of data sharing going on, all possible because N2K is open to
    all manufacturers.

No. It is partially open to members of the "Good Ol' Boys Club", and ONLY to those members. More to the point, it is certainly NOT "freely available to the general public", by whatever meaning of the term "freely" you care to apply.

Posted by: NotYetJoined in reply to Ben at April 4, 2012 2:40 AM

Speaking as one formerly involved with 0183 and 2K, and now the data comm standards world....IMO the answer to the question is "yes".

NMEA ... drop the spec fees and make up the difference on certification.

Given the relatively small ecosystem, 2K took an astonishingly long time to get to market. The effort was started right about the time the Web gained traction in mid 90s.

Posted by: Paul at April 4, 2012 11:30 AM

So anything that involves software should be "free", in both the "free beer" and "freedom" contexts, notyetjoined?

Thats nice of you, deciding what everyone else should do with their work.

"publicly available" doesn't mean that they have to give it away. No matter how much you plead poverty.

Posted by: BrianM in reply to NotYetJoined at April 4, 2012 5:14 PM

I generally agree with notyetjoined definition of an open standard. His presentation is consistent with accepted definitions across multiple industries.

You somehow made the leap that an open standard means there is no opportunity for NMEA to profit / fund the effort. Not so. My company makes massive profits building products around open standards, and a good chunk of that goes to supporting standards deveopment bodies with $$$ and people. That model might work for NMEA. Another approach could be to simply lower the spec costs and raise the certification costs. There are multiple ways this could be structured if NMEA was of a mind to lower the barrier to developing with 2k.

Its NMEAs game, they set the rules. But I can't call this an openly available standard, and I maintain its prevented the 2k ecosystem from arriving sooner with more product options.

Posted by: Paul in reply to BrianM at April 4, 2012 10:21 PM

I tried really hard not to get sucked into this thread :(

Ben...

"N2K is, in fact, the result of thousands of hours of largely volunteer labor contributed by members of a marine electronics trade organization"

Are you telling me these volunteers had no financial interest or are not employees of organizations with a financial interest in the 2k ecosystems? I'm sure a lot of very bright, hard working people contributed (I knew several of them). Passionate about the success of 2k ... sure. Volunteers ... without compensation? No

Posted by: Paul at April 4, 2012 10:35 PM

Quoting BrianM...

    So anything that involves software should be "free", in both
    the "free beer" and "freedom" contexts, notyetjoined?

I never said that, nor anything even remotely resembling that. It would appear that you are confabulating "open standard" vs. "open source". They are two different things; and we are discussing the former, not the latter. But as a side note, I'll point out that a lot of people make a lot of money with open-source software; for just one example: http://www.google.com/finance?client=ob&q=NYSE:RHT -- based on today's close, their current market capitalization is something north of 11.7 *BILLION*.

    Thats nice of you, deciding what everyone else should do with
    their work.

Straw man. Again.

    "publicly available" doesn't mean that they have to give it
    away. No matter how much you plead poverty.

I'm not pleading poverty; and even if I were, that would be completely beside the point. The issue is Ben's characterization of NMEA 2000 as an "open standard" -- which it clearly is NOT. A secondary issue implicit in the discussion is whether or not NMEA 2000 should be an open standard, and specifically how this would affect the level of innovation seen in the various products available to the marine-electronics consumer.

Posted by: NotYetJoined in reply to BrianM at April 5, 2012 2:03 AM

"The issue is Ben's characterization of NMEA 2000 as an 'open standard' -- which it clearly is NOT."

Actually it's NMEA that characterizes 2000 as an open standard. If I thought it wasn't, I'd say so; my interest is truth. But I started researching the definition of an "open standard" years ago, when this accusation first surfaced. What I concluded was that the editors of Wikipedia got it right. While there are several narrow definitions of "open standard" -- some of which are held to be THE definition with religious fervor -- the core characteristic is "publicly available" and NMEA 2000 is clearly publicly available.

How "NotJoinedYet" (NJY) can read the same Wiki entry and come to the opposite conclusion is rather astounding, as it requires complete denial of the ITU definition of "open standard" (as well as the careful words of the Wikipedia editors).

http://en.wikipedia.org/wiki/Open_standard#ITU-T_definition

There are two particular aspects of the ITU definition that contradict NJT. One is its definition of "Publicly Available" which goes like this: "Easily available for implementation and use, at a reasonable price. Publication of the text of a standard by others is permitted only with the prior approval of the SDO." In other words, the ITU is fine with charging for standards documentation like NMEA and many other standards development organizations (SDOs) do, and also with copyright restrictions about the distribution of those standards.

And the ITU has more to say about the permissible relationship between an open standard and intellectual property rights (IPR): "IPRs essential to implement the standard to be licensed to all applicants on a worldwide, non-discriminatory basis, either (1) for free and under other reasonable terms and conditions or (2) on reasonable terms and conditions (which may include monetary compensation)." So the ITU also permits licensing or royalties within an open standard system. Obviously their main focus is that an open standard be available without discrimination.

Whereas ANYONE reading this can buy the NMEA 2000 standard or parts of it immediately and build a product using it that will only require a certification process to be sold in unlimited quantities -- without royalties -- N2K is an open standard by the ITU definition.

Also obvious is how respectful Wikipedia is of the ITU definition. For one thing they listed it first. For another, they made their own core definition simply this: "An open standard is a standard that is publicly available and has various rights to use associated with it..." Yes they go on to note the various bodies, governments included, that define open standard as open and free, but only as an illustration of how "There is no single definition and interpretations vary with usage."

Note that the ITU is the freaking International Telecommunications Union, a creation of the United Nations. Is there a bigger or more important standards making organization on earth? I don't know the answer to that, but ITU standards are certainly critical to a lot of multi-manufacturer technologies we expect to function well and competitively, and that definitely includes equipment boaters depend on. Let's also note that while not as explicit the IEC, ANSI, ISO, and IEEE all seem to endorse an open standards definition more along the ITU lines than the definitions that NotYetJoined chose to quote on this thread. There are also all the CANbus related standards I detailed above that charge for documentation and demand certification like NMEA does.

In other words, the organizations actually responsible for building stuff have a broader definition of open standard than some government agencies whose main concern is keeping their IT departments from getting burned. Or the scads of computer folk who think any open standards must be free and even downloadable right now (yes, we've had that pronouncement).

NotJoinedYet also mis-characterized the NMEA as an "old boy's club". The inference is that you have to be invited to join or, if you can apply, that someone will decide whether you're worthy. NOT TRUE! "NotJoinedYet" could become a member of NMEA with just an application and a credit card, no review. And he or she should join if he or she is actually interested in having a say in NMEA 2000 and other standards. Though he or she doesn't have to join NMEA to purchase and use the standard. Because it's publicly available.

But "NotJoinedYet" is obviously mixed up about what joining means anyway, being the first person in some seven years to confuse registering on Panbo with joining something. NJY, registering simply means validating an email address so you're less likely to be a spammer and can post comments immediately. That's it, all of it. And, by the way, you can use your real name whether you post comments as a registered or unregistered user, and it will make them more credible. Though I doubt anyone will buy your insistence that all microprocessors are computers no matter how you comment. It's just another abuse of language for the sake of ideology.

At any rate, these last few comments are why I was really hoping we could avoid another pointless discussion of "open standards" on this thread. "Open and free" is just a subset of "open" and therefore NMEA 2000 is open. If you think you know better than Wikipedia, ITU, and countless others on this subject, good luck. And whereas this thread has gotten quite ungainly anyway, this will be the last comment. Here's a Forum thread where NJT can have at it:

http://www.panbo.com/forum/2012/04/is-nmea-2000-an-open-standard.html

Posted by: Ben at April 8, 2012 10:26 PM