Why MySQL 5.1 Is Not GA Yet, and How You Can Help
Yesterday I had a good conversation with Monty Widenius (a MySQL founder) about MySQL 5.1. Specifically, about the fact that MySQL 5.1 is not a GA (generally available) release.
My impression, which was wrong, was that it was difficult getting critical mass to download 5.1 and use it simply because it was not a GA release yet. I thought the paradox of needing to have a certain amount of usage before release was the barrier.
That’s not the case at all.
The problem is that bugs in MySQL 5.1 are not being fixed. Part of the way a bug is deemed critical enough to be worked on is how many people are running into the bug.
I am of the ilk that “me too” types of posts are somewhat frustrating. When I find a bug, I try to find it on http://bugs.mysql.com. If it is there, I add any relevant information. However, I do not simply write “this bug affects me too” because I am annoyed by “me too” type posts.
However, Monty explained to me that if we ever want MySQL 5.1 to become GA, we need to review the bug list for MySQL 5.1 and for any bugs that you want fixed, put a “me too” in there. Something like “I would use MySQL 5.1 if this bug were fixed” is sufficient.
July 22nd, 2008 at 6:58 pm
[…] don’t know about “me too” types of bug replies, but before everyone goes to the bug database and starts saying “me too”, “this […]
July 22nd, 2008 at 7:30 pm
Bugs are found when software is used in “interesting” environments.
MySQL 5.1 has seen plenty of downloads I’m sure.
The issue is indeed the feedback cycle not functioning. If someone finds a problem, reports it, but it doesn’t see action, then this causes a negative chain reaction for this person’s participation.
They won’t bother filing more problem reports, they won’t bother downloading newer 5.1 versions, and they might not even bother continuing to use 5.1 at all if the problems are hindering their work.
July 22nd, 2008 at 7:35 pm
Unfortunately you do, on some level, run into the same paradox, albeit indirectly. Because fewer people are using the non-GA version, you have fewer people to notice/report these bugs. This in addition to having people like you who tend to avoid echoing existing reports.
July 22nd, 2008 at 9:50 pm
That’s an interesting way of doing things. Shouldn’t the internal team be reviewing the severity of the bug and fixing it based on that? I just can’t understand the way MySqlAB (and now Sun) works sometimes…
July 23rd, 2008 at 3:18 am
Wouldn’t it be easier to have a counter that people could increment simply by pressing a button that says “This bug affects me too”? Seeing a counter with a high value would be better than having to trawl through all those “me too” posts.
July 23rd, 2008 at 3:38 am
Which is completely crazy. A bug exists in a bug tracker to be fixed, not to be fixed once it reaches enough “me toos!”
I wonder about the quality of Linux (the kernel) if a similar prioritisation model was adopted…
July 23rd, 2008 at 5:08 pm
Mike and James — I think the issue is that the internal team can’t get a good sense of priority of which bugs to fix unless there’s some way to see whether it’s affecting one person or many.
Tony — yes, it would be much easier, but there isn’t currently one there, and we’d like 5.1 released as soon as possible.
July 24th, 2008 at 9:10 am
If the “me too” model is not working, at the least they should just start fixing bugs. Just showing that things are being fixed in the bug tracker alone could go a long way for inspiring confidence in MySQL 5.1, increasing the usage rates, and then as a result increasing the bug reports.
Fixing bugs based on the number of reports received for a bug would work, if a large amount of users were using the system, if you only have a few users you’re only going to get a few bug reports, even if the bug is major.