Re: Planning to split doc-rfc
mdz@debian.org (Matt Zimmerman) wrote on 25.02.01 in <[🔎] 20010225181538.A20784@alcor.net>:
> On Sun, Feb 25, 2001 at 10:46:00AM +0200, Kai Henningsen wrote:
>
> > nick@debian.org (Nicolás Lichtmaier) wrote on 25.02.01 in
> > <[🔎] 20010225004815.A31844@debian.org>:
> >
> > > > DRAFT STANDARDS doc-rfc-draft-std
> > > > EXPERIMENTAL doc-rfc-experimental
> > > > HISTORIC doc-rfc-historic
> > > > PROPOSED STANDARDS doc-rfc-proposed-std
> > > > STANDARDS doc-rfc-standard
> > > > - doc-rfc-misc
> > > > Comments? Better names? Should there be a dummy doc-rfc package to
> > > > pull in the others?
> > >
> > > The packaging should be tematic,
> >
> > Are *you* going through the ~3000 RFCs and making a list which ones belong
> > to each theme?
>
> Eh? They are already classified by the RFC editor/IETF. However, their
> list is a little different from the one above. It goes something like:
Uh, their list is exactly what I'm using in the above proposal.
Or to put it precisely, they have two different lists: what a RFC is
published as, and what it's current status is. The above is the "current
status" list. (And it's significantly less buggy now, after I exchanged
several emails with the RFC Editor guys.)
> - FYI
> - BCP (Best Current Practice)
Ah. Forgot about those two, as they're only in the other list, for some
reason. Yes, should have those, too. (Both are small lists.)
> From the above, it looks as if "misc" will actually include "informational"
> RFCs, so perhaps it should be renamed to reflect that.
No, misc was emphatically *not* meant as "any rfc not anywhere else". I
originally did not intend to package _all_ rfcs (just as I didn't before).
I'm now pondering if maybe I should have prio extra packages for 500-packs
of those, like someone suggested. Need to see how that works out in
practice, first.
> > (And what will you do about overlaps? There's a *lot* of that.)
> >
> > It's not as if I haven't tried to come up with such a scheme for years.
>
> Overlap should be easily solved with dependencies.
I really can't see how.
> Users may install a
> subset, say, BCPs, and the "Informational" package would depend on that one.
We were talking *thematic* subsets here. Neither of these are thematic
subsets. Mail or SNMP are (and overlap is stuff like RFC 2789 "Mail
monitoring MIB").
MfG Kai
Reply to: