Author: Mogens Larsen
Date: 09:20:48 09/07/00
Go up one level in this thread
On September 07, 2000 at 11:07:09, pavel wrote: >I agree fully with this, >beta-testing is not about playing a bunch of games between computer programs. That's not entirely true. It is one aspect of betatesting, ie. to check whether a program behaves correctly with different timecontrols or under different GUI's. The games can also be used to tweak the opening book, engine parameters or against specific programs. Of course 1000 games against the same program at one specific timecontrol isn't all that important. >If that so then the guys from the chessbase and rebel can do it themselves. I seriously doubt that. It would take a lot of hired personnel to check if every minor modification to a program is good or bad with some certainty. We're not talking Microsoft here. >there can be another reason for beta-testers (IMO)-----> >for example Christian Koch is a beta-tester for "Gandalf" so he is playing a lot >of games with gandalf against other strong chess playing programs. and then >posting them in public forums, as a result "promoting" the "gandalf" engine in a >way to public. This has nothing to do with finding bugs for the programmer. That depends on how the betatester uses the games afterwards. Personally, I play fewer games but I check the games with opening encylopedia's and other programs afterwards to try and find errors with my limited chess knowledge. And also the logfiles if I have the time, but not as often. In the case of "Gandalf", where there is no company involved, games are definately very important as the programmer can't do everything himself. But I agree that spreading the word is also very important. More important for some than others. Mogens.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.