On Thu, 2004-08-12 at 21:08, Jason Huebel wrote: > On Thursday 12 August 2004 7:52 pm, John Davis wrote: > > I like the major versioning, but do we really need the minor versions? I > > see that as a throwback to what we are doing now. The reason that I want > > to drop release-based versioning is to avoid the needless replication of > > data since the profiles between same year releases are so similar. > > > > BTW - we really should start a GLEP about this ;) > > Well, it's not that we have minor versions. Frankly there may be years where > the first major version is all we need. But specifying that a revision > number is the accepted minor version would simply eliminate confusion. I > don't see how we could get away from some type of minor version, considering > we had one significant profile change this year when we moved to xorg-x11. I > know that amd64 is planning to have at least one more significant change when > we move to gcc 3.4 in 2004.3. > > As far as the GLEP, I think we should finish hashing it out here, then write > it up. :-) This is a very constructive discussion so far. We could decide to > write the GLEP when things digress. ;-) > > So, I still think using the year as the major version, with a revision number > as the minor version (only for -r1 or greater) is a good way to go. But > you're the head releng dude, so it's up to you. But I think we should at > least get away from using .1, .2, etc in the profiles. I can do with a revision number as the minor version, as long as it is the exception, not the rule ;) I agree as well that we should get away from the .x stuff. I am fine with the proposal that we have and am ready to GLEP it if there are no further comments. Thanks for the constructive discussion :) Cheers, -- John Davis Gentoo Linux Developer ---- GnuPG Public Key: Fingerprint: 4F9E 41F6 D072 5C1A 636C 2D46 B92C 4823 E281 41BB