Test software on machines customers actually use - was Re: strangest systems I've sent email from
toby at telegraphics.com.au
Fri May 20 14:44:28 CDT 2016
On 2016-05-20 3:03 PM, Fred Cisin wrote:
>> I don't know if that was a specific market ploy based on Moore's Law,
> an actually quite smart move, . . .
>> or just the generally accepted practice of getting an initial version
>> with the API working any which way, then refactoring to improve
>> performance/correctness in later versions.
> For decades, I used to rant that the biggest problem with Microsoft
> software was that they treated their programmers "too well".
> That if Microsoft programmer had space problems, they would immediately
> replace his machine with one with more RAM and bigger drive, and he
> wouldn't learn to be memory or disk space efficient.
> That if his programs were too slow, that they would immediately replace
> his machine with a faster one, and he would never learn to write fast or
> efficient code.
This is still a huge problem that afflicts web development as much as it
did desktop development.
Devs should have down-specced machines (say, 5+ years old) or at the
very least, should be regularly testing on them.
> If there was ever a hardware problem, they would immediately replace the
> machine. Accordingly, Microsoft programmers NEVER actually experienced
> hardware issues, and had to IMAGINE what disk errors, etc. would be
> like, resulting in software that couldn't properly handle them when they
> happened. ...
> OK, so development should be targeted for next generation hardware.
> BUT, testing should be done with what is actually out in the real world.
> Grumpy Ol' Fred cisin at xenosoft.com
More information about the cctalk