Brian "Krow" Aker's Idle Thoughts - Debian, MySQL, and why am I working on this bug?

LJ has been running MySQL on Debian for quite a few years now.

The big problem with this is not that there was miscommunication, but that there was a clear revisionist interruptation of MySQL's own documentation. As stated on Slashdot, "we were told that 'Generic Linux' in MySQL's list of supported platforms means 'generic versions of the implementations listed above'; not support for Linux in general." The proper information was clearly communicated, all that sales had to do was honor what MySQL has posted on it's own website and honor the industry accepted definations for the terms. But somehow sales was able to redefine "Generic Linux" to mean something redundant with what was already "listed above." So the meaning of the term was already listed before, why even bother listing it again with a supposily redundant term? The excuse does not make sense! But once sales is willing to act on definations that do not make any logical sense and do not match accepted industry definations then *ALL* of the documentation on the MySQL website/support becomes a meaningless mess of words which can have any random defination at any given time. The trust needed to advocate getting MySQL's foot in the door is much easier to loose than it is to gain.

In my own situation, I work for a major University's Information Tech department in which I was advocating to make use of MySQL. There is already some unoffical use of MySQL taking place (not supported by the department's DBA team) and the department has worked with FOSS centric companies in the past to help lower it's TCO. When the price of Oracle licensing came up, I advocated a hybred enviroment where some applications where switched to using MySQL instead and backed with a MySQL enterprise support contract. At the CIO's demand, there was a feasibility study done which included the DBA team which was agressively looking for any excuse to continue to remain centralized on Oracle. When the DBA team submitted it's report back, it included details about MySQL that didn't seem to match what was stated that did not match the website. The DBA team claimed the report was unbias and the detailed explainations (or as I would prefer to call "redefinitions") where the results of talking to MySQL sales on a conference call and each member of the DBA team backed that the damning wording was from MySQL and not their own. I was always under the impression that the report was because of bias on the part of the DBA team. But I think I now better understand how the situation played out. The bottom line was that we could not trust MySQL to the same degree that we currently trust Oracle. Regardless of if that trust was based on fact or fiction did not matter. What sales provided to the DBA report was "the facts" for purposes of moving forward with advocating further use of MySQL. In my case, it killed us ever going forward.

Oddly enough, the same sales person got back to me a couple months later about a discounted training session in my area. When I brought up that why the purchase was killed and that Oracle sales bundles in Oracle training sessions for "free," there was no apology or further discount provided. Since the "discounted" price exceeded how much my supervisor could ever justify spending for a product we do not offically run, I wasn't able to go.

I still consider MySQL to be a better solution for several of our own application development projects and I am still willing to be an advocate for MySQL. But at the end of the day, no matter how much I try to be an advocate for MySQL, I can not make up for how much MySQL sales wants to shoot themselves in the foot.

I recently spoke with some MySQL sales staff:
http://scottharmon.livejournal.com/34070.html

I was wondering if you could shed some light on the frequency of the GPL source releases?

Well, I meant versioned source releases. Distributions with less manpower (for example, Slackware), tend to rely on upstream a bit more for new versions to come out when there is some major bug, or a security issue.

Ah, sounds good then. So there will still be source releases for at least major bugs, and for security issues.