How are PostgreSQL transactions more real than MySQL using the BerkeleyDB or InnoDB? I have to admit I'm not familiar with the differences, but I was under the impression that they were the same.
After perusing the InnoDB/MySQL docs, I can see that MySQL is finally catching
up. So I'll grant you that one, although it requires "strapping on" a completely
separate solution from an independent vendor underneath, and that scares me.
Also, I think some people might not agree that PostgreSQL is faster than MySQL.
Did you notice the date on that benchmark? It's using Pg 7.0.2. I wouldn't have
used anything prior to 7.1 either. Some more recent benchmarks put them
in the same ballpark, with various tests showing one or the other ahead.
And how is it cheaper? They're both free!
Purchase price = $0. Installation time costs "money" though. I've always had
to wrestle to install a MySQL distro from source. Pg installed trivially the first time. I was really amazed. And development time: with true views and subselects,
you can do some amazing things, including putting business rules directly in the
database, rather than having to code them in every application (and forgetting one,
oops). Subselects, views, triggers, stored procedures. You don't miss them until
you miss them, and then you wonder how you ever got along without them.
So let me say it this way. MySQL is fine for people who are moving up from DBM.
But once you've played with Oracle, Pg is the way to go. You sacrifice far too much
of "typical" database coding with MySQL. Maybe someday MySQL will have everything in SQL 92. But Pg has that today.
-- Randal L. Schwartz, Perl hacker
| [reply] |
