Finally starting to Happen: Freedom of OS to the Industrial World?

Ask permission for reproduction.  26-Mar-2005

In my Article #1 I mentioned Industrial applications would be slower to come out of the World of Windows Alone, since they're tied there strongly by vendor driver policies. Early signs that this is going to change are now visible. That is Good News.
Finally starting to happen: Freedom of OS choice to the Industrial world?
I saw a tiny magazine article today, expressing the release of NI-DAQmx Base for Linux platform. A closer look reveals that their offering is actually (at least titled) open source, and provides support for Mac OS X as well. Great News!
To those of you wondering what? NI-DAQ is the Industry leading API (Applications Programming Interface) for Data Acquisition devices (reading, writing bits and pieces, either binary or analog values). NI is shorthand for National Instruments, a company who practically created the marketplace -and terminology- for computer based instrumentation, and still controls that as Microsoft used to control the desktop workingplace. Now, can we continue? :)
While NI's offspring itself is more based on the Mac than on Windows, the latter is where the company has triumphed. Up to the point that Industrial automation appliers have been more or less locked with the OS, did they want it or not. The story goes like this: "there's no drivers for Linux" -> why? -> because there's no demand -> why? -> because everyone's using Windows -> why? -> because there's no drivers for Linux. You can loop that a dozen of times or so.. to get an idea of how the martketplace has been.
Now, by this little news, I believe it has changed.
NI-DAQmx Base is only the first part. I expect from now on NI to steadily open up its approaches to embrace Win32, Linux and Mac OS X more and more as alternative, equally supported, OS platforms. This is very wise for NI itself, since it blocks the chance that its competitors would use this liability to race past them. Now, they challenge themselves, which is good.
Also, NI is quite prepared for this. They already have LabView running on all the three OSes (and PocketPC, too). Likewise Apple, they get part of their income from software, part from hardware, probably more or less 50/50. At least, the two pillars support each other, and they are probably one of the few in the business that can do so. As with Apple, they are the Best. (really, really, objective valuation, forgive me! )
But, there's even more to the announcement, if we care to look close..
"Open source", they say.. Does that mean NI-DAQmx Base is truly open source, or is it just the terminology. Do they welcome the Community to work on the drivers, and the whole framework? If they do so, and do it honestly, they're reacting to things like the Comedi driver set on Linux. Comedi provides open source data acquisition drivers for Linux, with support for any vendor's cards. NI-DAQmx Base, naturally, would support only NI's. Or at least, support NI's the best.
What next? If you have the drivers, what do you need to make a system work? Glue code.
LabView is the First Child of NI, and will remain to be their crown jewel. It won't be open sourced, at least not completely. I don't think that would serve their purposes. So.. what if people start glueing their code using Python, LuaX, or some other scripting solution instead of going LabView. Would that pose a threat?
With a different angle of observation, would NI support or just "watch by" such developments? Would they themselves actively parttake in that? After all, they do have the other foot in hardware sales, so even losing a bit in LabView licensing wouldn't hurt that much. And there's the known lack of good Development IDE's for scripting languages. Perhaps that would be a throw-in of sorts to using commercial NI tools. I do believe so.
The loser in this Game of Opening is to be TestStand. It's overly complex, very, very bound to the Win32 COM approach, and NI must by now be struggling with it themselves. I expect that product to be slowly declining in focus, rather than becoming more vital, and eventually experiencing the faith of the Dinosaurs, seeing smaller mammals (scripting languages of today) pass by and eat their prey. They simply don't have an ecological box to live in, any more.
Since the Industry marketplace is generally slow to react (Linux started its head run around 1999, Industry seems to be changing its course now 5 years later) this decline & replacement of TestStand will take some time, of course. Perhaps another 5 years, perhaps more. And NI will be selling licenses while this happens.
However, as an advice to NI (I do genuinly like them ) they could realize the situation themselves, and spend the TestStand development money on making the framework multi-OS, suitable for controlling (and debugging) scriptlets. Providing a free, basic, teaser version for evaluation and small scale applying, and the Full Suite for gaining money from the Big Boys. This would allow TestStand to reinvent itself, becoming slimmer and faster in the process.
Probably the New TestStand wouldn't share that much with the existing Win32 offering, but it would still fill the particular market place, just doing it in a new way. I wish NI is on this course, and if they are, I wish them all the luck on it. They do deserve to be around even in the 2010's. I hope they are.
- Asko Kauppi
Wishes for NI: - support your USB and FireWire GPIB products natively on Mac OS X. Currently, Ethernet, PCI and possibly PCMCIA adapters have OS X support.

Asko's Web