Comments on: Mechanism, not Policy

By: Walls

Walls — Wed, 25 Jul 2007 10:46:34 +0000

An excellent post, well done.

By: Tony

Tony — Wed, 25 Jul 2007 10:12:09 +0000

A great post put succinctly. It’s a point I’ve laboured for many years, which has for the most part fallen on deaf ears.

By: Maciej Makulski

Maciej Makulski — Wed, 25 Jul 2007 09:56:48 +0000

I’d say that comparing them all always comes down to the developer personal preference. They are products and they’re ruled by the same rules as any other products. They’ve been born because their was a requirement for them and then their users made them be different and then the competition changed them and will change them again so the differences between their capabilities will blur. The beauty of the process is that they will improve in every way to some point where they will meet the same SQL standard in the most efficient way but I seriously doubt that it would stop developers from „religious wars“ anyway, those war come from the natural human need to be accepted as part of the group/community and that will last as long as the human kind itself.
But that’s just my 5 cents. It’s a nice blog.

By: Mark Dennehy

Mark Dennehy — Wed, 25 Jul 2007 08:12:08 +0000

I know what you mean Keith, and I don’t disagree with it. To be fair, you have to admit there is the point of view of the vendors who’d consider the database and the hardware it’s running on as a single black box – but we both know that’s not a very good point of view 😀

Thing is I was more thinking about the way MySQL does things when I wrote this and not so much about the Oracle/Postgresql benchmark setup. It is definitely time someone set up a proper comparison benchmark between Postgres and MySQL though…

By: Keith Murphy

Keith Murphy — Wed, 25 Jul 2007 01:29:07 +0000

Thanks for the comment on my blog. My point wasn’t Oracle vs mysql vs postgresql. I try my hardest to stay out of religious wars 🙂 My point is that it isn’t fair to perform any comparison test when the hardware isn’t the same. It skews the testing no matter what is running on top of it. I wouldn’t even consider running Oracle (and I was an Oracle DBA for a while back in the late ’90’s) so it really doesn’t matter one whit to me. It just boils my blood when companies sell this as some kind of test when it isn’t. It is running two databases on two different sets of hardware. That is all it is. And I don’t know who put this test together, but if it was one of the vendors and not some independent (and I do mean independent..not bought and paid for by one of vendors) testing entity I would consider that highly deceptive.