Andrei POPESCU wrote: > Stretch and Buster were only meant as example, please consider my > nitpick to have been: > > The name of the next release is generally known, latest by the time > of the freeze of the previous one.
Sure. :-) > In this particular case the Release Team was slightly late, but they > made up for it by announcing the name also for the next release, which > makes the lives of a few package maintainers easier. I am curious. How does knowing the names further in advance make some things easier? For my part I think it confusing that a release candidate will have a lifecycle of two years (give or take). During that lifecycle it will start out identical to the previous and then grow into the new. We will have two years of email talking about Jessie and all of that will be about something different than the Jessie that releases next year. For example two years ago if you had installed Jessie you would have gotten sysvinit and today if you install Jessie you will get systemd. That both of those are facts make naming troublesome. Later when searching the archives for a problem in Jessie we find all manor of postings that talk about Jessie but are actually referring to quite a different set of bits. It isn't practical to do so because the name gets embedded in various places but in an ideal world I would dream of a process where the release would always be referred to as $release or something and then upon the day of freeze and release it then was given its name. Then we would have a "Jessie" that once named would then refer to the release from that point forward. Not practical to implement this in any way I can imagine but I can still dream about it because it would be better than we have now. Bob
signature.asc
Description: Digital signature
