Author: Robert Hyatt
Date: 10:44:20 09/06/00
Go up one level in this thread
On September 06, 2000 at 04:39:45, Uri Blass wrote: >On September 06, 2000 at 01:03:29, Adrien Regimbald wrote: > >>Hello, >> >>I saw the huge thread on beta-testing, and while I didn't read the whole thing >>(it quickly turned into petty insult matches), I feel I should point out some >>misconceptions. >> >>1. A beta tester is not a "real tester". A beta tester simply tries the major >>features of the program and suggests improvements, reports bugs, etc. A "real >>tester" is actually part of the development team, and in fact, sometimes this >>phase of devlopment in a serious software project can occupy more time than >>actually coding the whole thing. These testers systematically try to break the >>program - doing things that no-one would normally do, verifying outputs for test >>inputs using various techniques - black/white box testing, morph testing, >>recursion testing, etc, etc. Beta testers do this as well. I have done many. I received instructions on what to test, and had to answer lots of questions about how this worked, how that worked, how easy was the interface to use/understand, etc... Beta testing is usually done by someone outside the design / implementation team, to get a "fresh look". But it could be done by the design team and is in some cases... It isn't uncommon to have the design and implementation team develop the code, then turn it over the test/clean-up group <damn netscape to hell, it posted without me doing a thing but typing> to get it ready to ship. Then the design team gets another crack at it to see what the clean-up group did... etc. > >I admit that I do not understand the meaning of these techniques. > >> These testers are advanced coders themselves, and >>are highly paid for their work. >> >>2. Beta testers usually do not contribute significantly to the program - they >>point out small bugs, make suggestions, etc - mostly cosmetic issues and >>smoothing out unexpected problems with the GUI or perhaps a major functionality. This is true, but they _are_ traditionally paid for their services. Most often by being given a production copy of whatever they tested, assuming it is of real value. Other times they are simply paid a cash stipend for each beat test cycle. > >I admit that I am not a real tester by your definition. > >> Suggesting that a beta tester be paid anything for the 20 or 30 hours they >>spend "tinkering" (read - playing) with a program when perhaps a whole team of >>highly trained professionals for most likely a period of years over long and >>grueling hours is absurd in the extreme. > >I think that there are cases that it is more than 20-30 hours. > >Beta testers may have some understanding about the structure of chess programs. > >Examples: > >Beta testers can suggest ideas to improve the evaluation function of a program >based on watching many games(for example by suggesting changes to the piece >square table). >They may suggest ideas which kind of extensions to add to the program. >They may suggest ideas for a better time allocation(I found that the time >allocation of most programs do not follow the simple rule use more time when the >time control is slower and most programs use more time for the first move in 2 >hours/game and not in 2 hours/60 moves). > >Uri
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.