Randolph Chung <[EMAIL PROTECTED]> writes: > When we were working on potato boot-floppies, it was my understanding > that potato freeze != boot-floppies freeze. Several important additions > (http-fetch, etc) went in after the freeze.
I think http fetch was in before. Some other stuff trickled in later. > If this is the case for woody as well, then we should find out from > ajt when he thinks the installer portion of woody should be > feature-frozen. > Some thoughts on general project management and such... I happen to > think we do have a chance of having a basic installer system going by, > say, end of Jan 2001. One of the things that we might try to do more > explicitly with debian-installer is to try to set concrete milestones. I > think Joey already had a rough outline of this -- having udpkg and a > menu "driver" first, followed by debconf, etc etc etc. This doesn't need > to be an immutable "document" or such, but it'll give people an idea of > what they can help with, and help us track when we think the installer > will be ready. You're suggesting a similar approach to the Potato freeze, but Anthony has told me his plan is have a much shorter freeze cycle. If we're going to use debian-installer for woody, and we want the freeze to be much shorter, then you need to go *into* the freeze, already having a functioning installation system on all arches supported in Potato. You guys are really going to need the freeze period primarily for testing support and bug fixing. Sure, some features may creep in during freeze. But I think it would be a huge mistake to consider the freeze period as part (half) of your implementation period. -- .....Adam Di [EMAIL PROTECTED]<URL:http://www.onShore.com/>
