ietf-nntp HDR parameter proposal

Russ Allbery rra at stanford.edu
Mon Mar 31 14:27:30 PST 2003


Charles Lindsey <chl at clw.cs.man.ac.uk> writes:
> "Clive D.W. Feather" <clive at demon.net> writes:

>> I'm suggesting that the commonest case after "anything" and "just the
>> base headers" is "just the base headers plus Xref", so either we have a
>> package containing only Xref, or we have a package for base+Xref. Or
>> both, of course.

> That is the commonest form of the overview.fmt, but I am sure it will
> not be commonest usage of the HDR command.

Well, the most common usage will likely be HDR ALL, but next to that, I'm
fairly sure that it will be.  Why don't you think it would be?  The
easiest way of implementing HDR from a database is to use the overview
database, at which point why not provide everything that's in the overview
database?

> Why should anyone particularly want to use HDR on a variant header like
> Xref?

Why go to extra work to exclude it when it's no work to include it?

> Anyway, if we allow the OVERVIEW parameter (with the implication that it
> really does the same thing as the OVER command), then any problem with
> Xref just goes away.

I don't think including an OVERVIEW parameter without requiring LIST
OVERVIEW.FMT is a good idea for all the same reasons that LIST
OVERVIEW.FMT is problematic.  You have no way of knowing, for a particular
article, what headers are included in the overview, and looking at the
results of an OVER command for that article doesn't help.

> And where parameters take fixed values, then they are just being used as
> keywords, and I feel that good general practice is that keywords should
> look like words. Funny glyphs should be used for funny things. And the
> meaning of "ALL" is pretty obvious even if you do not read the document
> to see what it means.

Yeah, this is basically the way I feel too.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>



More information about the ietf-nntp mailing list