ietf-nntp HDR

[Russ Allbery]

Charles Lindsey <chl at clw.cs.man.ac.uk> writes:

> Could we get away with using "bytecount" and "linecount" as the official
> designations in LIST OFERVIEW.FMT? That would be incompatible with
> existing usage, but who or what would actually notice the change?

> That would lead naturally to solution (c). Not compatible with existing
> XHDR usage, but then (c) is not compatible with existing usage however
> it is done.

I don't understand what you think LIST OVERVIEW.FMT has to do with the HDR
command.  The commands are completely unrelated.

> But having HDR Lines not return the content of a genuine Lines header
> (if present) is just too ugly to comtemplate.

So do we want HDR to be able to retrieve non-header information?  That's
the fundamental question, isn't it?

If we do, then we need some way of distinguishing between headers and
non-headers, since no matter what we call the line count, someone may use
a header by that name, and the NNTP protocol should not assume a specific
article format.  That sounds really, really ugly to me.

The decision would be easier to make except for the fact that the Lines
header in the actual article is completely worthless and no normal reader
client actually wants to see it even if it thinks it does.

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