Correcting Cohen

Stuart Cohen, CEO of the Open Source Development Labs, wrote an editorial in BusinessWeek recently -- Sun's Big Open-Source Bet  -- that contained several inaccuracies about OpenSolaris. Stephen and Patrick responded to Cohen, but I thought I'd add my comments to correct Cohen as well.

First ...

Last year, Sun made its flagship operating system -- Solaris -- available as open source. Sort of. You see, Sun wrote its own open-source license. It's a license that many in the open-source community don't like, and with good reason. 


OpenSolaris is not "sort of" open source. It's open source as outlined by the Open Source Initiative (OSI) -- which was literally a core requirement for everyone on the project team before we even began opening the Solaris code. The license in question is the Common Development and Distribution License (CDDL). It was a new license last year, but Sun didn't go off in a corner and write it from scratch; instead, we modified an existing, already successful, already OSI-approved license -- the Mozilla Public License (MPL). The CDDL is not Sun's "own" open source license. Anyone can use it. It's a template license that can actually help consolidate many of the MPL derivatives that have developed over the years. Additionally, if "many in the open source community" don't like CDDL, as Cohen suggests, wouldn't it also logically follow that those same individuals have a problem with the MPL too? After all, the CDDL and the MPL are really very much alike (see redline diffs at the CDDL link above). Perhaps they do, but I generally don't hear very many people calling Mozilla's code "sort of" open source. Do you? Now, was the CDDL controversial with some in the open source community when it came out? Yes it was, but that's another issue altogether and has nothing to do with the fact that OpenSolaris is open source. For more on the license discussion from last year, you can see these links: here, here, here, here, here, here, here, here, here, here, here, here, here.

Next ...

Unlike with Linux, all the rights to any changes to the source code for Solaris go back to Sun. So any developers contributing to Solaris are literally working for Sun for free.


If you change an existing CDDL file, your changes go back to the community -- not Sun. If you write a new file and license it under CDDL, it's community code -- not Sun's. If you write your own proprietary file and link that file to a community CDDL file, your proprietary file is yours -- not Sun's. CDDL code is shared in an open commons, which includes non-Sun developers mixing rather freely with Sun engineers. If people were "working for Sun for free" then how does the following list make any sense: SchilliX, Nexenta, BeleniX, marTux, the PowerPC port, the DTrace port to BSD, the ZFS port to FUSE/Linux and DragonFly BSD?

Next ...

In my experience, people will work for free when they see that work as contributing to the greater common good -- but not to the bottom line of a global computing vendor. This part of Sun's strategy escapes me.


That's not our strategy. I'm sure Sun's true strategy "escapes" Cohen, but that's only because he doesn't understand it. This is really just a rhetorical trap where a position is mischaracterized and then called into question. It's a common technique practiced for thousands of years and easily seen for what it is --  a unsubstantiated misrepresentation.

Next ...

Time will tell if Schwartz can build a viable software ecosystem and vibrant development community around this approach.


OpenSolaris is already a "vibrant development community" in its very first year, which is a huge accomplishment. We can demonstrate this quite clearly by contributions (code, documentation, scripts), open conversations that reach hundreds of thousands, thousands of participants, user groups, communities, projects, millions of lines of source code from multiple Solaris Consolidations, a published roadmap, non-Sun distributions, an open development process (draft), a published Charter, an open governance (draft, draft), and ports (DTrace, ZFS, PowerPC). And there is a lot of infrastructure coming that will help encourage even more community participation. I'd say that the OpenSolaris community has already built a fine foundation in its first year. Wouldn't you?

That's it.


I'd be very interested if he is bothered to respond.


Posted by Alan Hargreaves on May 31, 2006 at 12:59 AM JST #

Hey, Alan. I'm not holding my breath. :)

Posted by Jim Grisanzio on May 31, 2006 at 01:26 AM JST #

And I wonder if any crud was because Sun isn't currently a member of OSDL? It does seem rather un-educated, and what I've sadly come to expect of OSDL recently...

Posted by Glynn on May 31, 2006 at 06:19 AM JST #

This only tells me that they are scared, its the same FUD tatics that MS has beening doing for years. This is the thing that gets me about the Linux lot, they don't seem to be learning from history. They seem to lose the fact that they can through all this crap to get where they are, so why can't anybody else do it? Are they that special?

The solaris community is a odd one, it probably has more in common with the apple loons than linux. Why because they sit through all this crap and more, but still stick with it. No matter how many time Sun has tried to kill it off were still here, still producing apps and howtos.

best case was with x86 solaris, sun did know how many there were till it pulled it. Its other thing about the solaris community its a bit to quite, we don't have all the zealots shouting on street courners, were too busy getting the job done.

Posted by kangcool on May 31, 2006 at 02:16 PM JST #

Rock on Jim. Rock on.

Posted by Jason on June 01, 2006 at 03:45 AM JST #

Post a Comment:
Comments are closed for this entry.


« July 2016

No bookmarks in folder