"Jeremy M. Praay" <jer### [at] questsoftwarecmo> wrote in message
news:47b1fd2b$1@news.povray.org...

> It's a good list, but it's rather sad that this isn't common-sense, isn't
> it?  :-)
>

Tell me about it. Half the devs don't even know how to do performance tests
on their code. I ask them how long a query takes to run and they look at me
like I was speaking klingon.

Best one was late last year. I'm reviewing one guy's code and a view that
he's written concerns me. So I run the query on the dev server, in the DB he
was using. I watch it for around 30 sec, then go to get coffee. It was still
running when I got back. (Total 8 min or so)

I call him over and tell him that it's totally unacceptable, he needs to go
and review the query. He comes back not 5 mon later and tell me, with a big
smile, that the view now rund in 5 seconds.

I run the new query and, at about the 5 sec mark, the first 20 or so rows
appear in the result pane. The query carries on executing. At 3 min, I call
the guy over and tell him there's still a problem.

"But," he says, "on my machine it finished in 5 seconds" I go to his
machine. He runs the query. The first few rows return in about 6 seconds.
"There!" he says, pointing at the still executing query "See, it's done"

<sigh>

Post a reply to this message