[cctalk] Re: Turbo Pascal Kermit for CP/M

On Mon, 8 Apr 2024, Bill Gunshannon via cctalk wrote: Presumably, it worked on the machine that they were using. Not everybody tests everything on all possible configurations. SOME companies don't test anything on any configuration other than what they are using. For example, at Microsoft, if a programmer has a hardware problem, they immediately replace the machine. Good working conditions. BUT, that means that they have substantially less experience than we do with problematic machines. Hence, disasters, such as SMARTDRV originally defaulting to having write cacheing ON, which gives data loss (blamed on other components!) EVERY time there is an unrecovered disk error, OR the user turns off the machine (normal practice for most users in the 1980s) before all the buffers are flushed. Proper testing would have found that before release. Beta testers of Windows 3.11 who reported that problem were told, "You have a HARDWARE problem. That's not our responsibility." Those who suggested that the OS at least had a responsibility of reporting and making graceful exit in such conditions were dropped from the Beta program for all future products. -- Grumpy Ol' Fred cisin(a)xenosoft.com