Stories (RSS)
Is OpenSolaris “Going Nowhere?”
(3 votes, average: 3.67 out of 5)
Loading ...
A colleague pointed out an interview with Sun’s Chief Open Source Officer, Simon Phipps, where he apparently responds to comments made by our CEO, (transcript) Ron Hovsepian, in a podcast with the Linux Foundation.
First, I think that the the crux of the issue that Ron expressed is a general frustration with Sun’s interaction with the Linux community. Many in the Linux community see OpenSolaris as too little and too late — an effort that would have been significant and possibly worth joining in on before Linux had matured into the enterprise class operating system that it is today.
While Novell, Red Hat, Canonical, Debian, and many others are working together on Linux as the platform of the future, we have Sun off trying to build its own community around an operating system that it has more or less exclusive control over. I have spoken to quite a few community members over the years that think that this is a distraction to “the mission” of promoting Linux, at best, and actively hostile to Linux at worst.
In my opinion, the industry has chosen to standardize on Linux as the Unix-type OS of the future, and Sun’s failure to join the party is unfortunate. We need a platform that is not only open source, but one that is not owned or controlled by any single player.
Sun had a choice of licenses when it released OpenSolaris, and opted to go with a licensing strategy that is incompatible with collaboration with the Linux kernel and many other projects. This is why many see Sun’s approach as trying to balance its proprietary and open source strategies. It’s licensing policies allow it to continue a proprietary line and development model, while also offering code as open source.
Sun’s policies regarding OpenOffice.org and code contributions have also hindered the project and limited contribution from outside sources. That might be why most Linux distros ship the ooo-build of OpenOffice.org, rather than Sun’s.
I would also like to see Simon or Sun back up the assertion that Sun is “the single largest contributor” to code on an “average” Linux disk. I think that the GNU Project, for one, might disagree with that one. The assertion is probably based on the size of OpenOffice.org, which is very large in terms of source code and is certainly a significant application — but also statistically misleading, really.
I don’t know where Novell stands in contributions to Linux overall, though I know we’re second in known Linux kernel contributions recently, and we make significant contributions to GNOME, KDE, OpenOffice.org (though Sun won’t accept some of them…), and many other open source projects.
I’m happy to give Sun credit where due — they do contribute a great deal to open source projects like GNOME, and of course they have done great work with OpenOffice.org. So has Novell, and I don’t think we get the credit we deserve, either.
As to the question of OpenSolaris and where it’s going. I think it’d be unfair to say that OpenSolaris is “going nowhere,” but it’s tempting to see it as the open source equivalent of Ralph Nader. A distraction that may just ensure that the wrong platform succeeds in the end.
> I don’t know where Novell stands in contributions to Linux overall
An of course incomplete but decent list of some of these at: http://en.opensuse.org/Novell_Supported_Projects
I’ll bet that your next post will be “Microsoft ‘Going Nowhere’?” or “Novell, How to give moral lessons when we have none”. Attacking the ones that contribute to open source and free software while making deals, support and act as one with those that whose only want to harm open source and free software. Right. Check. Thank you for your lesson.
I think you hit the nail on the head about Sun Zonker…but much of the same can be said about Novell. Novell has a software catalog that could forklift replace Microsoft in an IT environment and yet it chooses not to GPL most of it. I seriously believe that if Novell opened some of it’s products like GroupWise and eDirectory and gave them to the Linux community it would be a major turning point in Linux adoption. In a couple years we won’t talk about this option anymore because there will be mature and full featured GPL based messaging and directory systems in the Linux community…and they won’t be Novell products. I spoke with Ron about this at Gwavacon this year and I think more people need to do the same because although SUSE is the best enterprise class distro on the market today, it probably won’t always be and Novell has so much more to offer than YAD (Yet Another Distro).
Of course Solaris is a distraction to any “mission” surrounding Linux - it’s its own system, after all, and a rather good one, at that. Sun also used to have a very strict policy on certain issues where Linux is lacking (to choose the nice word), eg. API and ABI stability, or justification of development (you get grilled on why you want to replace an interface with another one, it doesn’t seem to be that way for Linux (the platform, not the kernel), where new components replace old ones every 3 months). That seems to change however.
As for licensing, the CDDL is a quite good license, and well balanced - the “play with noone (but rob the weaker)” ideal of the GPL is getting in the way every now and then, see the problems it has when you try to combine it with CDDL code, for an example.
What Sun really has to learn is to let go, and give away control - that’s the Achilles heel of OpenSolaris right now, and it’s hard for third parties to grab it (so Sun would have to learn the hard way, similar to gcc and egcs) because of the sheer manpower behind the Sun branch of the project. That last aspect is probably true for every once-closed-now-open project by a large company (that isn’t effectively abandonware, that is)
I’m not Simon or Sun but this study shows that Sun is by far the biggest corporate contributor to Debian. Those numbers don’t seem to include OO.o