CVI on Linux - a discussion from CVI mailing list

[Jumper_Bones]

Hello NI-CVI support team:

I submitted these comments below to the CVI mailing list just the other day. Another member of the CVI mailing list thought that my comments were worthy enough to be posted here as well. Hope this might start more discussion in this area. I agree with the other poster from last week that who said that, rather than wait for users to ask about a Linux port to CVI, that instead you might want to solicit them for their opinions, especially on the "how" and the "why" they plan to use CVI on Linux. I think you might find their replies are important, regardless if you never get CVI ported to Linux... Here is what I wrote:

--------------------------------------------

It doesn't seem that long ago that CVI was once ported to Sun SPARC, I forget if that was CVI 4 or 5 though (I never had a reason to use it, so I have no idea how good it was, or whether it's intent of allowing cross-platform CVI project development worked well). So you would think therefore that migrating the CVI IDE source to a port to Linux on SPARC or Linux on x86 would not be such an enormous effort...

I once had a brief chat with National Instruments' Shelley Erickson (sp?) at a local Measurement Studio seminar in Boston last Fall regarding CVI's future with respect to the soon to be released Measurement Studio for Visual Studio .NET. Shelley is the Measurement Studio product manager, I believe... She indicated at that time that CVI would most likely be broken off from the new Measurement Studio release bundle, which I might guess is coming out later this summer (?). This would make sense to me due to the "managed code" restrictions on .NET objects...

I mention this because you would think that this might allow CVI to lend itself even better to a future port to Linux, assuming that it's current duty to remain a Microsoft-based IDE as part of a larger bundle would now be reduced. I'll bet that there were once some high-level, behind-closed-door agreements between NI and Microsoft regarding Measurement Studio's future as it was being developed over these many years. Maybe now if CVI is taken out of that bundle it would allow it to chase other markets of opportunity...

[ This discussion is pointless of course, if you believe that Microsoft's .NET platform might eventually emulate the Sun J2EE-platform concept of allowing your .NET applications to run on any OS architecture someday. I won't be holding my breath until I see the .NET framework ported to something else other than Win32-based operating systems... But if there ever is one (especially if there's a .NET framework port on Linux), it would make a more convincing argument to follow NI's new "Measurement Studio for .NET" route. ]

So of course, I cast my vote for CVI on Linux. Besides, there are many more "grunt" C-programmers living in the trenches on Linux than on any other OS, and Linux is more of a world-wide accepted OS, probably more so than Microsoft OS's. It makes me think CVI would be a huge hit on Linux, especially if it was priced very, very aggressively.

Can you imagine the quantity of automation projects that would suddenly appear on SourceForge it they ever did port it to Linux and made it very cheap to own? Can you imagine the huge boost it would give to existing automated-based projects that are already ongoing on SourceForge? I would think the Open Source community would embrace CVI quite a bit, even if CVI's core remained closed, because CVI's open driver architecture and their historical overall intent of supporting open development in C. Besides, other than Metrowerks, I do not know of many good C-based IDE's on Linux in the first place, so maybe if CVI got there early it would not have much competition...

I would think it would be a win-win situation for both National Instruments and the automation development community in so many ways if they made a port of CVI to Linux.


I wonder if any of you out there have similar opinions... If so, get the word out...