When MySQL / Sun announced the other day that some advanced features of future versions of MySQL would only be made available in the enterprise (read: for-pay) edition of the product, people began fulminating openly about Sun's commitment to open source.  The MySQL situation itself isn't anywhere nearly as dire as it might sound, but that doesn't make people bristle any less.

For one, the company's only talking about advanced features that aren't even implemented in the product yet.  My feeling is that those features also are only likely to appeal to people who are paying for MySQL support in the first place -- and I'd give it a matter of months before an open-source shop writes its own MIT/GPL-licensed front-end for those same APIs.  (The features in question make use of APIs that are available in all versions of the product, not just the enterprise edition.)

That doesn't make this strategy any less egregious to some, even if MySQL had decided to do this long before Sun stepped in.  There are many ways to make open source into a commercial venture, and creating separate for-free and for-pay editions of a given open source product is only one way.  It's also not always the way that is going to be greeted with open arms, especially if you're bifurcating a product that always has existed in only one, wholly open implementation.

From what I can tell, it's not the move itself that has people upset, but the way it's been pitched to the MySQL community.  How you spring this sort of thing on people is sometimes even more important than the substance of it.  Doing things by fiat, no matter what the actual content of the message, implies that you cannot be trusted.

For perspective, I got on the phone and talked to Deb Woods, VP of product management at Ingres, one of the biggest alternatives to MySQL.  She wanted to underscore Ingres's full commitment to open source -- that it would offer the entire range of its products as open source offerings without resorting to any bifurcation between community vs. enterprise versions.

Deb was most specifically worried about the fact that the enterprise-level features would then get tested that much less broadly.  "If you do that," she said, "then you're not really getting the full benefits of the open source development model.  You want as many people as possible to get involved."

She also felt that bifurcating an open source product would take away the audience's credibility for the company.  "The more you bifurcate your product, the more you fork it; you start to tear away your community and weaken it.  If you're a contributor, you may ask yourself, 'Why do I want to get involved and contribute if they may not open up my contributions?'  There are a lot of ways to make money with open source without bifurcating the product."  One of the scenarios she went on to describe was hardware appliances, something that MySQL's new parent company knows a great deal about.

I still have high hopes for MySQL in Sun's hands, and I don't foresee the core product being made into something crippled or substandard.  But one of the future lessons for any open source company that wants to move to making money via multiple versions of its product is how to break that news to its community.  Maybe companies need open source diplomats as much as they need open source evangelists.