On 05 Aug 2001 16:16:15 +0100, Richard George wrote:
Lets's just recap here... first, it was Darren that named MySQL, not
me, so I
have no idea why I'm getting all the flack.
For sure, sorry, attributions often get lost in a flame-fest!
But, you line yourself up nicely for a good healthy proportion of the flack:
Secondly, for my application, MySQL works fine
...
The day may come when the site grows enough that transactions,
subselects
etc are of use to me
What we're trying to do is save people the bother of extensive rewrites to support a new database. Yes, DBI gives you database abstraction, but MySQL takes it all away again by forcing you to write line after line of code that compensates for it's limitations.
Also, the "the day may come" argument is dangerously seductive. There are monster bugs in MySQL *today* that make it a nightmare choice. My personal favourite was asking it to perform: UPDATE foo SET bar=bar+1 WHERE id=?; ... this would regularly hang the MySQL thread. We're not talking rocket science here, and yet I was forced to move trivial SQL tasks into my perl code. Of course, YMMV - this was on a table of 1.5m records.
I don't think you'll find that "everybody agrees" on anything, be it which OS, editor, distro, or mouse to use.
I think it was a case of "everybody at Martyn's workplace agrees..."
Personally, I'd far rather make the wrong choice of my own volition, than blindly follow what someone's told me is the "One True Way".
Fair comment. The other bonus of the Open Source ethos though is that many people have often done what you are trying to do before, and there's an opportunity to save time and effort by learning from their mistakes. I wouldn't be evangelising Postgres today if it wasn't for MJ Ray and Paul Russell beating me violently across the back of the head with a copy of it whenever I grumbled about MySQL crashing...
Andrew.