Things I Love About Transcript Episode 19: Why I rarely benchmark anymore In the olden days 4GLs were slow. In order to get any reasonable speed out them you had to carefully eek out the best performance from every handler.
So I used to benchmark various syntax options, eventually giving rise to HyperBench, SuperBench, and RevBench, tools to help me run those tests just a little easier. Today I was writing a function that I intuitively felt might be suboptimal, but since it's only called infrequently I was inclined to leave it alone as long as it worked. Well, habit got the better of me and I ran it through RevBench: it takes 0.16 milliseconds on average. Sure, I might be able to optimize it, but why bother? :) That's something I find with the engine a lot, that testing the speed of algorithms is not nearly as critical as with other 4GLs. I still test complex routines now and then, but most of the time I'm less surprised by discovering speed differences among syntax options than I am simply impressed by how fast all them run in Scott Raney's engine. -- Richard Gaskin Fourth World Media Corporation Developer of WebMerge: Publish any database on any Web site ___________________________________________________________ [EMAIL PROTECTED] http://www.FourthWorld.com Tel: 323-225-3717 AIM: FourthWorldInc _______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
