RANT: Never enough room

We did test, and test and test and test. We had databases filled with bug reports, thousands upon thousands of them for any given product. It's not that the test process doesn't exist or that it's not effective. It's the priority given to quality vs. ship dates and new feature sets. -- Ian ________________________________________ From: cctalk-bounces at classiccmp.org [cctalk-bounces at classiccmp.org] On Behalf Of Jim Stephens [jws at jwsss.com] Sent: Tuesday, September 20, 2011 3:16 PM To: On-Topic and Off-Topic Posts Subject: Re: RANT: Never enough room On 9/20/2011 2:06 PM, Al Kossow wrote: I would submit that the programmers should not be seeing flaky hardware upon which to do their development. The process should be to develop requirements, design, implement and test. The process at MS clearly had not found out how to test, and if the product was not handling the problem, then they failed totally up front in developing requirements. Since they rushed out a hack to rip of Stac-er or whatever it was, clearly almost one of the above methodology was followed, as you pointed out with illustrations. However it does no good to have a workstation in front of a programmer which doesn't work. The way to handle the problem is to give the problem of anticipating failures to someone who can develop test cases, which the product can be thrown at by the testers. Otherwise, how do you know if you have tested it again. One should be able to trace all the functions of the program back through to a use case which has been thought out, including ones which should be covering failing cases. You'd be way further down the line with that product than either what is there, or one developed by a programmer on flaky hardware, or unknown failure cases. the Beta program should not necessarily include someone with the problems you had w/o being sure they are doing the use case testing that makes sense, and in any case they should never ignore a concise report of a problem. jim