48-hour Lenovo Sale eCode CNET48DEC4
April 17, 2008 1:06 PM PDT
Random thoughts on the MySQL furor
Watching the MySQL uproar unfold brings to mind an array of random thoughts:
- Microsoft really should acquire an open-source company. It would get credit for opening up, even if it closed off some parts of whatever project it acquired. Sun...? It's getting lambasted for no logical reason at all. It's not as if MySQL wasn't actively considering tweaks to its model before Sun acquired them, just as all open-source companies do. No one has settled on the exact right model yet.
- While Microsoft and other "proprietary companies" move toward opening up, it would appear that there is some movement among "open-source companies" to close off. Maybe we'll meet in the middle?
- Marten Mickos suggests that he's not convinced Red Hat is getting fair value for its software (and, by extension, neither is MySQL). $1 billion seems like pretty good value for sub-$100 million in sales and scads of free downloads, but maybe I'm missing something. :-)
Of course, Marten is referring to ongoing sales value for the company, but it's unclear it would have ever managed to get the sales it has if it were proprietary. My guess is it would be the "M" in the "Maybe next time" stack without open source driving its popularity. So is it the case that's not getting paid fairly for its contributions or that it's only getting paid at all because of its open-source fan base? - There's a false belief that making software proprietary is the only way to get (serious) money from it. This is fueled by a convenient neglect of both past and future. The past was mostly a world of tangible goods and services around them. The future is all about letting free software and free services with clever ways of profiting therefrom.
In the past two or three decades we've had a Gold Rush of sorts - an "in-between" period when the old world of property met the new world of digitization. Software profited by trying to bolt the old world's property laws onto the new world's "property," but it's turning out to be a fool's hope. Software is not meant to be proprietary in the way that a car is. Its cost of production and distribution simply won't permit this errant attempt for long, as the entertainment industry is discovering with file-"sharing."
Don't get me wrong. We'll have many years of Oracle and IBM consolidating every last vendor on the planet in an attempt to prolong the Gold Rush. But eventually CIOs will en masse discover that they've been pillaged for two decades too long and will demand fair value for their coin. Time to embrace the future. - MySQL is getting more stick than most because it is a central piece of the open-source community puzzle. It doesn't have the same freedom to experiment that a Zimbra or SugarCRM might have. When it experiments, it affects tens to hundreds of millions of others. With that said, I don't begrudge its attempts to experiment.
- Speaking of experiments, here's one for you, MySQL: Create an AGPL-like license that will force all your freeloaders like Google and Digg to pay for the tremendous value they derive from MySQL or actually contribute fair value back in terms of code. Imagine how much more money you'd be making if Google would stop cheerleading for open source and instead start paying for it. If it's making billions distributing MySQL-based services over the web, exact a small toll.
- It's somewhat ironic that MySQL is moving in this direction after Jonathan Schwartz has been saying for the past two years that enterprises pay for mission-critical deployments of open source. I suppose it's likely that MySQL will get paid more once it starts selling seriously into the enterprise, rather than to the Web 2.0 freeloaders. Sun should be of help here.
Anyway, it's unfortunate that MySQL is being put under the spotlight. Perhaps if it more clearly separated out its Community product from its Enterprise product, the backlash wouldn't have been as sharp. But maybe not: People complain because MySQL matters. If MySQL were a rubbish company and project no one would care.
Matt Asay is general manager of the Americas and vice president of business development at Alfresco, and has nearly a decade of operational experience with commercial open source and regularly speaks and publishes on open-source business strategy. He is a member of the CNET Blog Network and is not an employee of CNET. Disclosure.
- Topics:
-
Industry news,
-
Strategy
- Tags:
-
MySQL,
-
open source,
-
community
- Bookmark:
- Digg
- Del.icio.us
Paul
That is why the OSS model is so powerful.
In order to "fork" the code base, you need an all-new set of developers, as the likelihood is vanishingly small that Sun would permit their employees to join what amounts to a rebellion against them.
When Firebird forked off of Interbase, it took a fair period of time for it to become clear that the fork was viable, and that came with the very positive factor that the "rebels" included the original author. I don't think it's terribly likely that many of the option-encumbered senior MySQL developers would be prepared to make the same sort of "leap."
Furthermore, if a fork took place, there would be a *huge* question as to which branch of OurSQL would be concentrated on. Would it be:
- Something akin to v3.23, in that that's what so many web apps use?
- Or something like 4.0? Or 4.1? Or 5.0? Or 5.1? Note that if you have *limited* development resources, as there's not the horde of 100 developers paid to work full time on it, it doesn't make sense to divide those resources across trying to support four (or more) fairly wildly different database systems.
- Similarly, the set of supportable storage engines would be likely to need to shrink, as the respective licensing deals between MySQL AB and Oracle (InnoDB) and IBM (SolidDB) won't apply to the "OurSQL rebels", nor will it be likely that Sun/MySQL will be particularly cooperative concerning licensing of Falcon.
Project forks can and do take place, but they always come at *considerable* cost, and the above details that the "rebellion" that would take place with MySQL(tm) would not trivially transform into a coherent successful community project.