Oracle, MySQL and the GPL: don't take Monty's word for it
In the continuing soap opera of Oracle's battle against the European Commission for the right to acquire Sun, and with it, MySQL, we have had to rely on the bloggers and analysts to get it right, because the media surely has not. Before you read any further, stop right now and read Matthew Aslett's excellent summary of Oracle-MySQL through last week, Pamela Jones' excellent piece on the matter (and her later update), and Matt Asay's highlighting of Monty Widenius' conflict of interest in opposing the Sun acquisition.
One of the more damaging consequences of this case is the opportunistic piling on of the GPL license. Monty, apparently, now hates the GPL - the license that MySQL chose so they could sell licenses for the proprietary version. And now, every BSD Tom, Dick and Harry with an axe to grind about Richard Stallman, the GPL, and GNU has stepped up to the plate, on cue, to deliver unsubstantiated rants against the GPL. Oh yeah, and don't forget that Linux kicked the s*** out of BSD and some folks are still sore about that. And in the ultimate irony, Richard Stallman himself joined the fray against... Richard Stallman?
The first thing that the anti-GPL FUD brigade latches onto is that the GPL effectively prevents forking a project. This is baseless. As we all know, trademark law doesn't just apply to the GPL. If a company owns the trademark for X, you can't just sell a different version of X and call it that. Red Hat has aggressively used trademark law to pursue Red Hat clones, as has a number of other companies that sell open source products. What the GPL *does* forbid, however, is forking code you don't own the copyright for and then issuing the forked project under another license, which effectively kills any outside attempts at the dual license strategy championed by MySQL and now Oracle.
A BSD-style license wouldn't frown on someone taking someone else's code and incorporating it into their products, as long as they comply with the stipulations that accompany those licenses. The GPL expressly forbids the dual license approach, but it does not in any way, shape or form prevent someone from forking MySQL. Matt Asay and others have argued that this effectively kills any forking of GPL projects, but you'll note that the trademark law, mentioned above, already does that anyway, regardless of license. Quick, name any commercially-backed open source project that was forked and eclipsed the original project - Joomla - Mambo is the one that comes immediately to mind, which is licensed under the GPL2, thus negating that argument. Whether issued under reciprocal licenses a la the GPL or more liberal licenses, such as BSD and Apache, a successful fork is a pretty rare event. If the GPL more successfully prevented forks, then there would surely be empirical evidence of such. At this stage, the BSD-style advocates want us to take that as an article of faith, but I'm not buying it until someone can show me actual evidence.
Now who would oppose an action that prevented the dual licensing of MySQL by a party that's not Oracle? Could it be someone who has forked the MySQL code and is now looking to use the very same dual license strategy? Now guess who's forked MySQL and could really use the permission to dual license it. Any guesses? Surprise! It's Monty Widenius!
So let's see if I've got this right - there's another company that stands to lose money if Oracle is allowed to utilize the GPL2 to its fullest extent and has now petitioned the European Commission to stop the acquisition because... Monty's European and the EC should protect their own? Seriously, if someone can show me a scenario that doesn't involve Monty being hopelessly compromised, opportunistic and self-serving, you're welcome to point that out in the comments. Until then, I recommend you follow the money, or in this case, the potential money.
One would think that paid journalists would have been on the ball and pointed out Monty's inherent conflict of interest, but that's expecting too much. I wasn't even going to comment on this case, because the people I respect have gotten it right, except that the technology press continues to get it so wrong. Thank goodness that informed bloggers are around to set the story straight.
By the way, I'm not suggesting that Oracle is great, and that there's nothing to fear from them. I'm only pointing out that those leading the anti-acquisition campaign are resorting to underhanded tactics and not disclosing their own bias. It's not OK to say we're for a specific license only when the company that utilizes said license passes our own personal smell test. That smacks of cronyism and flies in the face of what open source is supposed to be.