Aaron J. Seigo wrote: >-----BEGIN PGP SIGNED MESSAGE----- >Hash: SHA1 > >On Tuesday 15 October 2002 09:00, Kevin Anderson wrote: > >>MySQL is like MS Access. It's an OK database, but it doesn't scale nearly >>as far as PostgreSQL. >> > >very true. > >>MySQL doesn't support many features such as transactions or views. >> > >well, it goes beyond just lacking features. > >insert into myTable set someRow = 'value'; > >say WHAT? and the inane rationalizations that come out of the MySQL devels to >explain their lack of features is sometimes maddening. "referential integrity >logic doesn't belong in the database back end". what? > >>MySQL designates simplicity over functionality. >> > >simplicity? not really. it advocates a non-standard way of doing things over >standard, more powerful and safer mechanisms that aren't really more complex. >it does advocate simplicity (in how it handles data) over safety and in the >hopes of achieving greater speed (which it doesn't except for some very >specific cases). > >just because you have the ability to do subselects doesn't mean you have to >use them. but when you want to or need to, having them is a good thing. > >>MySQL is MUCH easier to find information about, especially in the PHP >>realm, leading me to also believe it is more popular there. >> > >either that or: > > o MySQL needs more documentation because it doesn't follow standards > o the docs that comes with pgsql are enough > o the people who tend to pick pgsql know what they are doing > >;-) > >>MySQL's lacking features may well not matter for this installation. >> > >yes, for things like web boards it hardly matters. > >>Converting from one to the other isn't the end of the world. >> > >it isn't when you design your app around MySQL's kludgy implementation of SQL >and then realize that to achieve real performance you need to rewrite the lot >of it. > >ask the dbmail guys. > >>Overall, this is a geek version of the "Dodge rules, Ford is crap" >>argument. >> > >no it isn't. MySQL is not an ACID RDBMS while PostreSQL is. there is a >measurable and real difference as far as data safety and performance >scalability goes. this isn't a matter of taste, it's a matter of differences >in design where one doesn't care about serious work or adhering to standards. > >- -- >Aaron J. Seigo >GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA EE75 D6B7 2EB1 A7F1 DB43 > >"Everything should be made as simple as possible, but not simpler" > - Albert Einstein >-----BEGIN PGP SIGNATURE----- >Version: GnuPG v1.0.7 (GNU/Linux) > >iD8DBQE9rIRQ1rcusafx20MRAuqsAJ93X1ZMNmwkZHMJX7Z125H0FocqrQCfaU+U >dNRQCdW35K7DF7SZ+uFrY/o= >=IHxM >-----END PGP SIGNATURE----- > > Just to add to what Aaron said about mysql... I had a "simple database" project that I used my university textbooks to design and normalize to Boyce-Codd normal form before looking at what the limitations were in mysql. Mysql supported neither intersects nor subselects which forced me to de-normalize my design.
Had I been using a fully relational database I would have been able to use a correct design. Without these capabilities, I was forced to use a design which is going to be difficult to modify and will not scale well.
