On Thu, 2008-01-17 at 12:12 +0800, Tito Mari Francis Escaño wrote: > Sun acquiring MySQL? There's no love lost here. Feature for feature, > PostgreSQL is still better. > MySQL is like that <snip>
I'm not so doctrinaire. I'm a postgresql bigot (as ian says), but I do nuance. Although in any mysql vs postgresql discussion I almost always side with postgresql (except for certain very high volume, non-ACID requiring scenarios where postgresql's MVCC implementation gets in the way) a pragmatist. It's a good thing for mysql to be bought by Sun, IMO. After grabbing market share by being simple and fast, MySQL finally is moving in the direction of correctness (although, too slowly for me, mysql will remain a toy for perhaps a year or two more). Sun will, I hope, give it a boost in this direction. If MySQL becomes as fast, correct and reliable as postgresql at some point, this would be a good thing for both and for the open source community, generally. In general, mysql implements a lot of features that the market demands (and historically enterprisey features and, ACID wasn't a big demand). It has tended to implement hacks (some of which mysql regrets now) which tick the feature box but which are either sort-of-good-enough (most features) or (sometimes) weak, wrogn, brittle or dumb. for instance, for a long time you couldn't have a default be a function. partly that was because you couldn't even have user defined functions at all, but, e.g., dt date default now() would be good. since you couldn't have the default be a function call though, but mysql still had to implement default to current date or time, they implemented that "first timestamp with default null defaults to the current time" convention. It'll be hard to step away from that sort of bogosity. backward compatibility is always a challenge, arguably, one of the major problems with the microsoft is all the backward compatibility hacks that remain in the code, making maintenance and forward progress harder. MySQL will probably follow microsoft in this regard. They'll maintain bogus features for backward compatibility, but downplay them and implement a strict mode (very similar to perl's "use strict", which deactivates bogosities in that language), e.g., http://dev.mysql.com/doc/refman/5.0/en/server-sql-mode.html good lord, there's an ALLOW_INVALID_DATES option, which was the default before 5.0.2. As I said, mysql is moving in the right direction. it's still too bogus to use for anyone who wants their database to be correct. But that'll slowly get fixed, I'll look at mysql again a year or two from now. tiger _________________________________________________ Philippine Linux Users' Group (PLUG) Mailing List [email protected] (#PLUG @ irc.free.net.ph) Read the Guidelines: http://linux.org.ph/lists Searchable Archives: http://archives.free.net.ph
