Colin Watson wrote:
On Mon, Apr 01, 2002 at 02:49:53PM +0200, Fabio Massimo Di Nitto wrote:If we always have to stick 100% to Policies and possible breaks than we should just leave things like they are now. People rised a serious problem and the idea is to try to optimize as much as we can.Any complexity theorist will tell you that optimizing away constant factors gains you nothing in the long run, especially when you lose robustness by doing so. Given that the number of packages in Debian is growing more than linearly over time, breaking clean abstractions in the name of a small constant factor improvement that will barely last until the next release is foolish.
So if Im not mistaken you simply mean: leave things like they are, right? btw I don't see why trying to optimize things can make things less robust.
IF and only IF we can come to a real better techincal solution noonewill tell us not to change policies. It's obvious that during a transition phase things can break but unstable is there.. ;)The attitude that "unstable is there to be broken" is one of the reasons we have trouble making releases. What happened to professionalism and the art of designing good transition plans?
No please don't misunderstrand me at that level. what I mean if even with a good transition plan, an error can happen and we might need to accept that something can break and unstable is the testplan, isn't it?
Fabio -- To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org