Havoc's Blog

I think there’s some misunderstanding here. Several points:

• For FC4 specifically, there are some hard limits on the size of
  Core. So lots of stuff had to be moved to Extras – including many
  things besides AbiWord and Gnumeric – for practical reasons. This is
  just reality, and I don’t think it’s fair or reasonable to take it
  personally. If you suggest something else to remove instead, would it
  be fair for the developers of that something else to feel that you had
  a personal thing against them?
• Extras is not a “dumping ground,” it’s in CVS in the same way as Core,
  has the same quality guidelines as Core, and is easily accessible from
  yum in the same way as Core. Extras is simply the universe of all
  packages and Core is a subset. In other words, Extras is a part of
  the Fedora Project. The Core/Extras split is about how large a
  subset release engineering deals with when building the ISO
  images. And this decision to remove some packages is caused by release
  engineering and mirror problems if we go above 4 ISOs in FC4.
• Many, many people install a minimal system, reboot, and then “yum
  install” the remainder of the OS. (This was always the way with Debian
  back in the day as well, when you had to do this to avoid dselect
  doom.) For people doing this, Core vs. Extras basically doesn’t
  matter. Even things not in Extras are easy to use; I have jpackage in my yum.conf for example.
• The long-term plan (for FC5) will probably be to either move a lot
  more stuff to Extras, shrinking Core to be more or less “the
  default install only,” or to move the things we temporarily moved out
  for FC4 back in to Core. So this whole thing is just a short-term
  situation.
• The Red Hat desktop team is planning to keep maintaining the AbiWord
  and Gnumeric packages in Extras for now.

All this is explained in the huuuuuuuuuuuuuuge mailing list thread
about what to remove; but I can’t really blame anyone if they didn’t
read the whole thing. Anyone is welcome to read it and suggest
something else to remove instead of the current list, but I think
anyone reading it will realize that the problem is intractable, and
requires the long-term plan of either a) re-expanding Core or b)
shrinking Core to be some relatively objective set like “the
default install.”

On the overall office suite question, the Red Hat engineering team is
very supportive of a native (and just plain better) office suite. I am
personally as well, and was recently re-convinced of the value of this
when I tried out Pages and Keynote. But given that we have to remove
something, it just makes sense to remove duplicate functionality
first, instead of for example removing all office suites in
order to keep a choice of ten text editors.

(This post was originally found at http://log.ometer.com/2005-03.html#2)