-- by Barnacules Nerdgasm
"https://www.youtube.com/watch?v=S9kn8_oztsA"
TL;DR (FR-version) :
- 1/ Ils ont licensié un département entier de testeurs quand ils ont fusionné leurs multiples systèmes d'exploitation : Desktop , Phone, etc.
- 2/ Ils ne testent plus du hardware réel; tout est virtualisé (i.e. les problèmes spécifiques au matériel ne sont pas révélés)
- 3/ Ils se reposent beaucoup sur le _Windows Insider Program_ et la télémétrie pour avoir des remontée de bugs avant de déployer les mises-à-jour "au plus grand nombre" (mais ce mécanisme a des limitations : les utilisateur du _Windows Insider Program_ ne remontent pas toujours les problèmes, les _mini-dumps_ ne sont pas toujours suffisants pour identifier la source d'un problème, les problèmes sont résolus à tâtons, et dans les pire cas, complètement ratés et les correctifs causent davantage de régressions).
TL;DR (EN-version):
- 1/ They did a massive lay-off and got rid of a whole testing departement when they merged their multiple Operating System teams : Desktop, Phone, etc.
- 2/ They don't test real hardware anymore; tests are run in virtual machines (ie. hardware-related corner cases are not tested)
- 3/ They rely a lot on the _Windows Insider Program_ and telemetry before rolling-out updates to "most people" (but this process has limitations : _Windows Insider Program_ users don't always report bugs, _mini-dumps_ aren't always enough for tracking down a bug, some problèmes are solved by groping and, in the worst cases, even missed entirely and attempts to fix a problème cause further regressions).



Cloning a VM that is part of a domain in order to use it simultaneously along with the original one is hard: on top of changing 1/ the MAC address, 2/ the computer name and 3/ the IP, we always run into mysterious problems when faced with the domain controller.
So, cloning a VM creates conflicts difficult to solve. Even if it has been taken off the domain prior to duplication! My network administrator pointed out that this is caused by the SID of the machines, but we never really knew what the solution was in order to work around this problem (from a user point of view, not from the network admin point of view). Now, I just found a possible solution in the Xen documents. The tool mentionned below (sysprep) is not specific to Xen, but to Microsoft. (Source: 