• E-mail this page
• |  Print this page
• | 
• | 
• | 

On MySQL's About-Face: It's About Expectations

The official word from Kaj Arnö of MySQL / Sun is out: Portions of MySQL that were originally being considered as closed-source components will now be open source as well.  Good news, bad news, or none of the above?  I take the third view.  The real issue is, again, not open vs. closed code, but how you engage the open source community -- how you clue them in to what kind of company you are.

Originally, I'd felt that MySQL was not making some terrible mistake by offering closed-source components as add-ons for MySQL -- just that the way it was being talked about with its audience and user base wasn't what it could be.  It was about expectations.  People didn't want to feel like MySQL was going to turn around and make arbitrary decisions about what was open and what was not.  They wanted clear marching signals; they wanted an inherent level of trust.  They wanted to know these guys were on their side and not the side of the robber barons, so to speak.

And then there were folks, like a software developer buddy of mine, who took another tack.  He was quite proud of the fact that MySQL was trying to monetize its work in a way that would be minimally intrusive for the majority of users.  The people most likely to need the higher-end features that MySQL was to charge for would pay for them anyway, and the APIs that exposed those features weren't going to be locked up, so what was the problem?  He hadn't expected a backstabbing, and in his purview he hadn't gotten one.

The other day I talked to an outfit that deploys a unified messaging solution (I'll have more to say about their product in detail later) with Linux and other FOSS products as the base, but with the product itself not being open source.  Their point of view about the whole thing was that it was a question of need.  Their product was meant to be deployed as a unified solution with a minimum of modification needed, and they have a full complement of developer's APIs and a free-to-use 20-user version of the product, so for them an open source version isn't really vital.  And as long as their prospective customers knew where they were coming from ahead of time, they felt there wouldn't be any issues about open vs. closed source in their case.

In the end, a lot of this comes down to two things: what the users expect from a company's offerings, and how that company lives up to those expectations in the public eye.  You're never going to completely get rid of the risk of being misinterpreted, but you can do things to minimize that risk.

« Does Your Top Management 'Get' IT? | Main | Alfresco's Social Computing Slant Shows ECM's Evolution »