He set the OpenSolaris community asunder. He caused the OpenSolaris and Indiana fiasco.

Just one thought about MySQL. The development model of it has evolved over the years. It began as a couple of individuals who put it together. Many early big features, like Windows or Replication, were put together or paid for by individuals in the community.

MySQL is “Organic” open source though. It has never been a closed source product. It is difficult, though not impossible, to take closed source and take it in the direction of open source. It took Mozilla a long time to become a Firefox.

I know that Monty, David, and I believe that we need to find a way to open up the development model of MySQL. We have recognized for the last few years that this is one of our weakness as a project.

Cheers,
-Brian

It is therefore hilarious to me, as it should be to all, that Sun calls itself an open source company. They have not got the foggiest clue, and this is all the more evident each 8-14 month period when I load a newer version of Ubuntu on the old box.

Linux improves because of trust. Period.

I think you and I are talking to completely different levels of scope in this discussion. I’m taking it far broader from a strategy POV and while you’re talking about incremental progress within the project, and while that may be ok for today…. Sun Tzu would say Sun has built a strategy that has already lost the war.

It seems that what you’re saying is that we shouldn’t have even attempted to open up Solaris, except perhaps to have thrown the code over the wall and let other projects pick over the remains. And if OpenSolaris was simply a nice collection of otherwise distinct technologies, I could almost see that point. However it’s not that but rather a much more cohesive thing which is why we believed at the time (and continue to believe) that it’s worth having a community around it.

Roy Fielding’s departure is unfortunate and I wish that we had his perspective as we move forward. That said, what I do know is there are a lot of passionate people about the technology, both within and outside of Sun, and the OpenSolaris community is a place for them to interact. And call me naive or hopelessly optimistic, but I do believe in the coming months that you will see the Mercurial repository outside the firewall and that more and more of the various processes that are currently awkwardly transitioning between being inside Sun and going outside will be more firmly outside.

I’m talking about a fundamental difference and approach that will not work for OpenSolaris – I’ve said it many times before and Sun is only proving my point. If Jonathan wanted XYZ to happen, he has the control to make it happen.

And I have to disagree with you on the license actually. I don’t believe there would be any difference if Sun had used the GPL for OpenSolaris. It’s not that easy to just mix/mash the codebases.

I know plenty about OpenSolaris. There was a time when I thought Sun was really going to open up and turn around. Then I peeked enough to find out the details. I suggest everyone start taking a critical eye to the nonsense I heard coming out of Sun executives’ mouths – take them to task on it and start to really ask the right questions. I’ll bet you don’t get a good answer.

Trademarks in this example are a very simple, known entity to work with. Sun is not blazing new ground having an open source project that it owns a trademark. Just ask anyone at DLA Piper down the road, and this trademark BS would be resolved in a couple weeks. Perhaps some should question if that’s really what Sun wants (I think Roy figured it out).

Yes, the source code for a large part of OpenSolaris *is* still behind the Sun firewall. That’s considered a bug by everyone in the community including everyone I know at Sun. It has taken *too* long to get where we are but the fact is (whether you wish to believe me or not) is that progress has been made. It’s been slow progress and folks like “stevel” (and others both inside and outside of Sun) have been instrumental in moving things forward. But you characterize the situation that Sun *wishes* to keep the source behind its firewall which is absolutely not the case.

Next, you know as well as I do why Sun requires joint copyright assignment. It’s exactly the same reason many open source projects require the same thing. Yes, Sun is a corporation and isn’t a non-for-profit consortium and hence might be considered “suspect” by some. And perhaps it would have been nice to have had some sort of not-for-profit consortium which had the copyright of the code that was open sourced and which joint copyright could be assigned for new code. But I suspect in order to do so it would have taken even longer to get where we are today.

Which brings me to the part of the binary tarball you refer to. As with getting the source outside the firewall, people at Sun would very much like to see the bits and pieces which are not open source “emancipated”. I wish this was a higher priority of the project but the fact is that many people in the community put more value on development for *new* drivers and the like, particularly since there are redistributable binary drivers available for the older components.

Coming from the Linux world, I realize this may seem strange and again, you may not believe Sun’s intentions here but if you take a look at the massive amount of change that has gone into OpenSolaris from a code perspective the past three years, practically all of that new development has been with code that is open source.

Tel, there are a number of reasons Sun decided to not use the GPL when it open sources Solaris but creating something GPL-incompatible was not one of them. Please see Simon Phipps many postings about this.

That said, you’re probably right that if Sun had used GPL in the first place, many of the innovations in Solaris would have been slurped up into Linux. And I think the world would not have been a better place for it. The fact is that Linux for all of its positive aspects is not the end all, be all of operating systems. It doesn’t solve all the problems that users and developers have and there appear to be some fundamental differences to the way the respective communities deal with certain technical challenges. That’s not to say that OpenSolaris always has it right but neither does Linux always have the upper hand. The fact that you’ve never even peeked at OpenSolaris means that it’s likely you’re ignorant of its strengths when compared to Linux.

To me, the main thing that convinved me never to even peek at OpenSolaris was the complexity of the legalities and licensing. Sun was re-issuing code under this license and that license, no one has the time or energy to keep up. From the outset, if Sun publishes under something GPL-compatible then the good code will get slurped straight into Linux and the rest will be dumped. So Sun published under something non-GPL compatible (I believe) and anyone working on Solaris ends up building a tiny island for themselves, right out of the mainstream with no way to get back other than start from scratch (and maybe not even).

The name of the project is irrelevant. CentOS and RHEL get along just fine, everyone knows the deal. If you build some experimental code and you build around CentOS while you are at the “starving artist” stage then after an IPO you can buy RHEL, port all your code across (including custom kernel modules, library hacks, whatever) and it all just works and RedHat are going to get onside for support. No bad feelings, no uncomfortable questions about legality, no BS, just get the job done.

Suppose you were building some experimental new filesystem. Would you base your work on OpenSolaris? Might as well put a noose around your own neck really. Call it Indiana, or call it George, the same problem happens. People doing groundbreaking work are going to either use Linux (if they are cool with the GPL) or use a BSD (if they don’t like GPL).