We should not care what "our competitor" is doing if it's not what we think is right.  Rolling out an update that breaks X is a serious error on Ubuntu's part.  It clearly doesn't happen often (I think this is the first time), but regardless, even one such case can shake the foundation of people's belief in its reliability, as it clearly has (read some blogs and the ubuntuforums).  The question is how do we learn from this and improve Ubuntu - both the process and the OS itself - with the goal of making sure that it doesn't happen again and that if it does, the effects can be reversed by the average user so that their computer is still usable until a fix is rolled out. So the way I see it thus far the discussion has taken two paths: 1) The process - there needs to be better screening of updates prior to getting released worldwide.  Some have suggested an alternative yum repository.  Others have noted that there already is a "dapper-proposed" repository that would allow testers to test out updates before they get rolled.  This seems like a very good way to get tests done on a decent sized number of systems and to get feedback before pushing out an update. 2) The software - there needs to be reasonable fallback mechanism in the case where a user installs an update that borks their system.  This should be rather straightforward using apt since that is the one officially supported mechanism for installing software.  You can already do this manually with an 'apt-get install package=version' , but it would be relatively simple to keep track of and revert the last updates made.  RedHat's up2date utility has a 'rollback' option. I tried to find a spec for number 2, but didn't.  Is there one that i'm missing?  Otherwise I'll create it myself. -SS- <div>On 8/23/06, Hervé Fache <<a href="mailto:Herve@lucidia.net" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">Herve@lucidia.net</a>> wrote:<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> On 8/23/06, Micah J. Cowan <<a href="mailto:micah@cowan.name" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">micah@cowan.name</a>> wrote: > > our competitor has already done this sort of things, and serious companies' > > IT do not roll updates until they feel it's safe to do so > > I don't think it's wise to distinguish between "serious" companies and > others. Serious was refering to the IT service, not the company (not obvious in English, I know). > And I can't think of any time in the remotely recent past when "our > competitor" has actually broken a system due to automatic upgrades... Come to my place and I'll show you a broken driver (official, from Windows update) hanging up XP, or rebooting it unexpectedly, depending on its humor! Fix: go to nVidia's site and get their driver. Would _any_ user think of that? Hope that clarifies, Hervé. -- ubuntu-devel mailing list <a href="mailto:ubuntu-devel@lists.ubuntu.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">ubuntu-devel@lists.ubuntu.com</a> <a href="https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)"> https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel </a> </blockquote></div>