Re: Base shrinkage proposal
> What you are proposing is basically to install partial packages --
> only the bits of packages we think we need.
Sort of; except that we will make sure the package managers *know* that the
packages are not the real/latest ones, so that they can be properly replaced.
> I understand your desire to do this, but I personally think that
> floppy installation is pretty much obsoletely anyway, and we should
> focus our efforts on making network and CDROM installation easier to
I do agree that network/CDROM installs are more common and we should make
sure those work well. However, I do not agree that floppy installs are such
a minority that we shouldn't really care. From what I have gathered, there
are still lots of people who do installs from floppies. Moreover, as I had
mentioned in my previous message, I think having 10+ install floppies
produces other problems (publicity, etc) that are bad for the project. Even
if it true that a small number of people use the base floppies for install,
people would look and *see* that we have 10+ base floppies; it's pretty
obvious what the ensuing reaction is.
> are PCMCIA laptops who can't use CDROMs. Seems like a pretty small
> sector -- our efforts would be better set towards having an
> alternative boot-floppies flavor that put the PCMCIA stuff on the root
That's a good idea too.
> The operative fact here is that once we start messing with base in
> terms of pulling out stuff, it will really take quite a while of
> testing to see what bugs and problems that may be introducing. I
> don't think it can be done prior to Jan 15th, the freeze date.
Again I don't quite agree. There's a lot of stuff we can pull out with no
bad effects. Just removing things like /usr/doc, /usr/share/doc,
/usr/share/man, /usr/share/info can bring significant savings with very
I do understand it is getting close to the freeze date and this seems like a
fairly major change, but I think it can be done with fairly minimal changes
Debian Developer <email@example.com>