Why 5.4?
The second most frequent question people have asked me since MySQLConf started is “Why 5.4? What happened to 5.2 and 5.3? Why not 5.5?” I got an answer to this at lunch today from someone who was involved in the decision making process on this. So here is the story as I understand it.
Why not 5.2? IF you recall your ancient history, 5.2 was the original plan for post-5.1 features, before it was renamed to 6.0 for marketing reasons. Because of some internal issues at Sun/MySQL and also to reduce confusion they did not want to reuse that same version number for the next product. This makes a lot of sense to me, reusing the same version number for two different products could definitely cause some confusion.
Why not 6.0? Short answer is that because of the reduced featureset, the new version didn’t warrant a whole new major version number. I think this is correct, because if they released 6.0 without Falcon, Maria, Online Backup, etc it would feel pretty strange. Instead we get the impression that this is a stepping stone on the way to the features that were originally promised in 6.0.
Why not 5.5? Well 5.5 can imply that you are “halfway” to something else, which again implies a major new featureset, and the 5.4 features are really incremental (though awesome* and much needed**) features. They can now save up the more impressive 5.5 number for a release that adds features more dramatically.
So why 5.4 instead of 5.3? The answer I heard on this was that they wanted to create a gap between the aborted 5.2 version and what comes next. Maybe someone else can clarify a bit better on this, but this was the answer I heard. I guess this makes sense too… if they skipped just one it would seem like they are missing something, if they skip two versions it’s more clear that it is just an arbitrary number to indicate the incremental improvements.
So when you add it all up, the version number 5.4 actually starts to make a bit more sense.
* awesome: performance on multicore so we can better scale vertically
** much needed: signal/resignal in stored procs
April 22nd, 2009 at 6:51 pm
Just a quick post to corroborate your story. I was sitting at the same lunch table listening to the same explanation about version 5.4’s name. I believe it was Trudy Pelzer who was explaining this to us.
April 22nd, 2009 at 7:04 pm
yup, I just didn’t want her to get the blame if I transcribed it wrong.