From: Russ Allbery (rra@stanford.edu)
Date: Fri Apr 13 2001 - 21:32:45 CDT
John Stanley <stanley@peak.org> writes:
> I think Russ is right, if I understand him right. X-headers are too
> nebulous for serious use, but they make arguments about what to call the
> header while it is still experimental go away. Maybe they should go away
> completely.
The problem that I see with X-headers is that people are too hard-line
about them, based on some (IMO) ill-conceived statements about them in
standards. If you're adding information that isn't ever going to be
standardized, is only there for purely informational purposes, is never
going to be interpreted in any special way by software, and isn't intended
to be part of any protocol, then *that's* when you should use an X-header.
An example would be X-Quote containing some automatically generated quote
or other similar kinds of vanity headers, an X-Poster-NNTP-Posting-Host
header added by a moderator containing the contents of NNTP-Posting-Host
as the article arrived in his inbox, or the X-Gateway header that I
referred to in the section on gatewaying that's only intended to be parsed
and understood by the software at one particular site.
Those are all good reasons to use an X-header.
What's arguably broken are statements along the lines of "any header not
standardized somewhere has to begin with X-". This may work for things
where there's an established IANA registry and publishing a full-blown RFC
for something isn't required to register it, or where adding a new type
may cause significant problems for existing software, but I don't think it
makes any real sense for news and it just makes it unnecessarily hard to
standardize new headers since you *have* to rename them in the process,
thus breaking all of the software that's already using them.
This is particularly an issue when the standards process is glacially
slow, as it historically has been for Usenet.
I don't think there's a clear-cut answer here; I can see the arguments
either way. But X-headers aren't a panacea by any stretch; forcing new
stuff into the X- namespace has some real problems.
> For example, some groups have been using keywords in the subject to help
> readers select articles. Are you telling me that a group that finds that
> the keywords "RE" or "SV" are appropriately descriptive should find
> those keywords stripped from its articles because an ignorant agent
> thought they were "back references", even though "back reference" has a
> simple and explicit definition that neither of those fits?
This is a very good point. I moderate one newsgroup that uses keywords at
the beginning of the subject line quite widely to tag posts as belonging
to one of the many fictional universes discussed in that group. Suppose
we had some universe that abbreviated well to SV? A good Subject line
would then be:
SV: Sylvan Vales #1
or the like, and if my news reader started stripping that SV: off because
it thought that was a mistranslated back-reference, I'd start getting
extremely annoyed.
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>