From: Russ Allbery (rra@stanford.edu)
Date: Tue Oct 08 2002 - 17:51:53 CDT
greg andruk <gja@meowing.net> writes:
> I believe that Brad's "kludge extra header" is the pragmatic thing to
> do. It preserves backward compatibility and lets us do the new things
> in a sane way. We can always leave in the "For your newsgroups file:"
> wart for legacy reasons, it doesn't really get in the way.
Yeah, that works. It at least works well enough to start writing code,
and we can always revisit later to turn it into something more formal.
I started playing around with the idea of a new newgroup control message
specification based on an extensible format sort of like what was
standardized in the multipart/report spec, and I think that has a lot of
potential and would make it much easier to experiment with group
attributes down the road. But that's also a whole additional can of worms
and would likely distract a lot of attention from the immediate problem
that we're trying to solve.
> checkgroups is definitely another story, it really does need to be done
> over from scratch to allow additional group attributes to be listed. The
> (Moderated) suffix thingy is cute, but there are limits to how cute one
> wants to get.
checkgroups have the potential to get *really* long, if you start
including named and extensible fields for each group. I don't have a good
feel for how much of a problem that might be.
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>