From: Austin Ziegler <halostatue@...>
Date: 2006-09-13T04:37:42+09:00
Subject: Re: Joel Spolsky on languages for web programming

On 9/12/06, William Grosso <wgrosso@wgrosso.com> wrote:
> Austin Ziegler wrote:
> > SQL databases suck, but MySQL implements SQL so badly that it
> > shouldn't even be called an SQL database.
> I think we're running into a terminology issue here. If we think
> of MySQL as a data persistence and retrieval mechanism, it has
> certain properties and is optimized for certain types of
> applications.

You mean "applications where the integrity of the data is completely
irrelevant."

> The fact that it's so widely used, and so effectively used by so
> many people, argues that the design decisions embodied in MySQl
> weren't completely wrong.

If you're trying to embody SQL database semantics, then said design
decisions are wrong. If you're just trying to have something that's
slightly more structured than flat files, then you're right -- they're
not wrong. Otherwise, MySQL is crap. Utter and complete crap for which
I have no respect for its developers.

> The fact that it's an incomplete realization of Codd's vision, and
> has some interesting flaws from a data integrity perspective, means
> that it isn't good for other applications.

Stop. Right. There. (1) *NO* current SQL-based database realizes
Codd's vision. Some come closer than others, but SQL databases aren't
properly relational. See the discussions headed by Fabian Pascal and
Chris Date (of dbdebunk) for that. (2) MySQL doesn't even implement
SQL database semantics properly. They *pretend* to, but they don't
actually do it. Because they don't support it, (3) a lot of developers
have assumed that if MySQL does it, it must be right. It isn't. It
isn't even *close* to right. I would argue that 90%+ of all the open
source database schema produced to work with MySQL take advantage of
MySQLisms and aren't even first normal form.

There's a place for big piles of data. But *not* in a SQL database or
a schema implemented on top of that database. I personally want RoR
databases to be *well-designed* databases, not the piles of utter crap
that you see from 99% of PHP developers who couldn't tell data
modelling from a hole in the ground.

Yes, I have very strong opinions about data modelling. No, I don't
think you can be a good developer if you don't know data modelling
(more provocative: you can't make a good object model if you can't do
proper data modelling). Yes, I think you can create good databases
that will run on MySQL -- but you have to do so *despite* MySQL's
stupidities.

> Generic statements like "MySQL is a big sticking pile that should be left
> in the scrapheap" aren't really useful at all.

When they're accurate -- as "MySQL is a big stinking pile.." is,
they're quite useful. If you want to work with databases, work with a
real database. Postgres, for example. If you just need something
quick, use SQLite. Both will protect your data better than MySQL.

It's unfortunate that MySQL has become the "preferred" database for
web work, because it's complete crap.

-austin
-- 
Austin Ziegler * halostatue@gmail.com * http://www.halostatue.ca/
               * austin@halostatue.ca * http://www.halostatue.ca/feed/
               * austin@zieglers.ca