Open Source Labview core and $ for services, add-on's....

[pavanb]

I did a search on this forum to see if the subject been asked before but couldn't find it quickly, so I post. I have been seeing business push and pressure to move in this direction over the years, more so recently and have migrated to eclipse for my current project and like CDT. I saw the following today and it got me off my inertia to ask/feel the question here. I would venture that Operating systems and other mature s/w tools aren't rocket science anymore, but commodities and open-source w/ customization makes sense. Easier said that done. Thoughts?

pavan

This inspired me to ask the question:  The Eclipse Helios simultaneous release of 39 Eclipse projects and 33 million lines of code showcases the diversity and innovation going on inside the Eclipse ecosystem. In terms of statistics, the Helios release includes 33 million lines of code developed by about 500 Eclipse.org committers from 44 companies. The important thing to remember about Helios and Eclipse simultaneous releases in general is that even though it's a simultaneous release, it doesn't mean these projects are unified. Each project is a separate open source project within Eclipse.org, operating with its own project leadership, its own committers and its own development plan. The simultaneous-release concept is designed to provide a transparent and predictable development cycle.

Other ref:

Microsoft ups the ante in the robotics market, makes MSRDS free ... (see coments from emiliekopp)

CentOS, Ubuntu...

Eclipse & addons

Python

GCC

Octave

Scilab

Android (I think it is open?)

Open-office

ISO formats for docs

Open Car

Open prosthetics

...

The list goes on.

[tst]

Are you talking about free or open source? According to the article, the MS announcment is about making the product free, not OS. Altenbach has already suggested a free hobby version here:

Noncommerc ial Hobby/Home license for LabVIEW

If you're talking about OS, there are a couple of differences between Eclipse and LV - first, Eclipse (if I understand correctly) is mainly an IDE for existing languages. In the case of LV, the product incorporates both the IDE and the language, which isn't standard. Opening up the IDE would require either exposing the language specification or creating APIs which will allow clean access to it. Seems like a lot of work for something that will probably provide relatively little gain. Altenbach's idea should probably be enough to draw in a larger crowd.

The second difference is that an OS business model might not necessarily fit NI as a company. Presumably, today NI uses LV and the hardware sales to drive each other up. If the IDE was OS, people might have less need to buy stuff from NI. Obviously, it is possible that going OS will drive NI's sales of hardware and support 20,000% up, but how likely is that to happen?

[pavanb]

After I posted, I remembered someone had suggested a free hobby version and yes the question is - free LV core first, OS IDE next, OS language (contributors from around the globe and/or crowd sourced/innocentive like addons).... Why shouldn't the marketplace desire/push for this? I know it is a loaded question and quite possibly impractical for a company, what about the industry and community, a consortium (firefox is an example)? I have seen the migration to free, OS... firsthand (some smooth some not) and wonder when, where and if the chasm will be crossed in the NI, Mathworks, Agilent... community. What will it do to our current mode of operation/the ecosystems biz model (sorry I couldn't resist the overuse of biz model etc). If it can be imagined, it is possible, maybe it isn't feasible or easy but when has that stopped things from being made and becoming a reality.

[pavanb]

I recently heard about an enterprise license which gives unlimited LV access to a individuals/comunity within its umbrella. For many new NI users benefiting from this it looks like "free" LV. An encouraging step towards a model of LV core being open some day, hopefully not too too far away.

---

tst wrote: it is possible that going OS will drive NI's sales of hardware and support 20,000% up, but how likely is that to happen?

---

I think the likely hood is there, depending upon how it is done. How high? I bet NI and the industry have talked about it years ago and continue to shift, adjust. This will affect NI in the next 10 years. "Free" LV seems to be one indication of this trend.

[JackDunaway]

Correct me if I'm wrong: Open Source seems to be a selling point in software markets that are already choked with competition. Open Source is the competitive edge that makes that product stand out from the rest - the customer realizes that Open Source has more flexibility than the proprietary competitors, and in most cases is free of charge (although support or advanced modules may be for sale...).

So, tell me this: who is LabVIEW's main competitor that is going to force LabVIEW to go Open Source to keep customer's interest piqued?

I have a tiny view of the big picture looking up from the bottom of the hierarchy, but it seems NI has no motivation (yet) to go Open Source...

Now on the other hand, I can certainly see current mutual value for the "free/low cost personal edition (with proprietary core source code)".

[pavanb]

It is a loaded question to a loaded topic, I'll try to answer but can't do justice. Competition? several, depends on the market/app. Simulink, C, Python, LV vs Testand vs Veristand, MS VPL, Need for more linux support, VITA/VME, UEI, open or more-open/strong h/w-platform is competition too, Scilab, Eclipse/gcc, Octave, Blender.... NI is aware of their competition for the most part, on disruptive technologies/movements it is hard to judge. Businesses are choosing not to migrate to NI/LV or/and are resisting expensive tools in this market and this benefits NI'scompetition, whatever their size or future. I have seen and am seeing this first hand in various forms more so in the recent past & now. What has your experience been?

[JackDunaway]

> It is a loaded question to a loaded topic

Yes, my question was a rhetorical exercise to generate discussion. Note: I'm more specifically talking about LabVIEW than NI's product offering as a whole.

> What has your experience been?

Let's just say, "I have a tiny view of the big picture looking up from the bottom of the hierarchy". Very generally speaking, NI touts LabVIEW as a turnkey, dummy-proof programming language that quickly fulfills your software/hardware requirements. This mantra manifests itself in the current product offerings: Express VIs, plenty of add-on Toolkits, a large Example Library of common, simple tasks, a rich cache of free Reference Architecture frameworks available for download... and rightly so, LabVIEW steps up to meet the quick-and-dirty needs of even the least capable programmers. But LabVIEW is not just an 'easy' cheap date.

It's powerful. There is another trend that has emerged - meeting the needs of the rockstar programmer. Look at LVOOP introduced a few years back, look at Scripting which just became officially supported last year, look at the open architecture supported by Quick Drop... this is a trend catering to a higher echelon of skilled developers, and LabVIEW continues to push it's envelope higher on the user skill continuum.

Even from my lowly vantage point, it's evident there are vocal ranks in a tug-of-war from within who battle between "easy to use" and "powerful". The Decisions Behind the Design is a good example where we are candidly told, "We have decided to limit your power in some respects to make the entire package more widely accessible." (I'm not griping here - this philosophy is probably why LabVIEW is such a good fit for me!)

Also, of the development environments you listed, none seem to have a close enough "fit, form, and function" to be direct threats to LabVIEW market share. What other development environment can turn a group of smart electrical engineers into a group of deft programmers within a few months?

Not to mention, a while back I suggested the entire vi.lib be "unlocked" with no password protection. The response was sub-tepid. Apparently, nobody really wanted access to the lower level.... high-level abstraction layers and API's seem to be kings in LabVIEW.

It seems sentiments at NI are just not yet compatible with enveloping the programming cognoscenti by releasing the LabVIEW core open source, and it seems there's not enough external or internal market pressure to require this bold move.

So, that's what my experience has been.

[JackDunaway]

Re-reading the thread, I notice you throw out a 10 year figure... This Idea as a whole is not unreasonable over a long period of time, such as 10 years. My comments are fairly short-sighted, just pointing out that NI is not currently primed or seemingly even aimed at such a move toward open source.

[tst]

> I recently heard about an enterprise license which gives unlimited LV access to a individuals/comunity within its umbrella. For many new NI users benefiting from this it looks like "free" LV. An encouraging step towards a model of LV core being open some day, hopefully not too too far away.

I'd just like to point out that this should not raise your expectations. A site license is simply another type of distribution license and fits in a standard business model. It does not indicate an intention of opening the source and you can find numerous programs with site licenses which are not OS.

[pavanb]

This was an enterprise license (more "open"/border than a site or one company) involving over 3,000 engineers across organizations, I realize it isn't free and was a special case, the first of its kind I hear. Every little counts towards a move to "open" (used loosely) is the point I was trying to make. I haven't read the following but thought about it in this context,

Free! Why $0.00 Is the Future of Business

The goal when I started this thread was to get a broader & diverse perspective from the audience what OS or low-cost or more open or free means to an individual and organization in different parts of the world.

[JackDunaway]

Open Source/Proprietary and Free/Paid are two orthogonal parameters. You can can have Free Proprietary, Free Open Source, Paid Proprietary, and the weird one is Paid Open Source (you pay us and we'll give you the source code if you sign an NDA and agree to not redistribute the source code). I think the LabVIEW market is primed for Free, but not Open Source.