Quote:
Originally Posted by HarryT
I'd be very surprised if there were to be a widespread problem with iPhone/iPad. They are such popular products that they must surely have been tested!
|
One would think, but Microsoft missed a problem with
webcams and Skype. They own Skype, it's very popular, and the Logitech webcams in question are among the most popular webcams to use with Skype. The problem actually affects far more programs than just Skype, but you'd
think they'd have tested their own software. But they didn't.
All signs are that Microsoft does nearly no testing on updates now, particularly for Windows 10. Back in 2014, Microsoft laid off nearly all their "programmatic testers" as part of a
massive cut of 18,000 employees. Programmatic testers are the ones that develop programs to test updates against. In other words they're quality assurance. After this Microsoft began depending on developers to code their own programmatic tests.
The problem with that approach is two-fold: 1. Developers tend to not spend nearly as much time on developing the tests as they do on coding. This is just human nature, you focus on your main task (coding new stuff) and less on the side-tasks (testing that new stuff). Then you have 2. Developers tend to have trouble finding their
own mistakes. This is also human nature, you coded it so you tend to think it works because
it works the way you intended it to be used. Bugs often only show up when things are used differently than developers expected them to be used. So it's important in all types of software development to have testers hammering on it that try things that aren't planned. It's been known for a
long time in software development that this is true.
It appears that Microsoft thinks the Windows Insider program is good enough to replace all the quality assurance in Windows 10, but signs are that it's not. There have been major bugs in the Anniversary Update that weren't caught by insiders. (Including serious
problems with Kindle devices, apropos to this thread.) There's also the fact that Microsoft is not handling the bug reports it does receive through the program very well. They seem to be using next to no actual people to go through the reports, and rely on scripts to decide what problems are reported most often. This is, quite obviously, going to miss a
lot of stuff.
So don't count on Microsoft testing things thoroughly, testing seems to be an afterthought to them now. Major tech publications are starting to
call them out on it. (Notably, the author of that Article, Peter Bright, is an unabashed Microsoft cheerleader. When you have even him criticizing Microsoft, something is
seriously wrong.)