[Top] [All Lists]

Re: New I-D: draft-koch-subject-tags-considered-00.txt

2004-11-19 20:12:29

 Peter Koch <pk(_at_)TechFak(_dot_)Uni-Bielefeld(_dot_)DE> writes:

       Title           : Subject: [tags] Considered Harmful
       Author(s)       : P. Koch
       Filename        : draft-koch-subject-tags-considered-00.txt
       Pages           : 8
       Date            : 2004-11-18
Various mailing lists modify the Subject: header field of messages
  sent to the list to contain a kind of identifier enclosed in square
  brackets.  Every now and then this 'feature' is also requested for
  the main IETF list.  This document collects pros and cons of this
  approach, tries to identify the real issue and will, in a later
  stage, try to give a recommendation.

A URL for this Internet-Draft is:

Hmmm! This is a minefield.

Generally speaking, meaningful tags in Subject fields (specifically any
which may need to be globally recognized by MUAs) are a Bad Thing. The
problem is that people regularly invent new conventions for same, usually
without full understanding of the impact elsewhere, and some of these
conventions then become widely used, and even find their way into

The best known example is the "Re: " hack, which is so widely used and
recognized now that there is no turning back from it. The problem is that
it has also spawned variants such as "Re[2]: ", "Re: Re: " and "Sv: ", and
nobody knows which ones are 'official' and which not. And yet software
needs to recognize at least the "Re: " form, for purposes of presenting
lists of messages in a sensible order (because you cannot rely on
threading by the References header field alone to do the whole job).

Much the same applies to "Fwd: " (which again has variant forms), and
there is am IMAP draft <draft-ietf-imapext-sort-17.txt> which seeks to
codify some arbitrary selection of such conventions (including some of
the form "[...]"). Personally, I think this is a Bad Thing. The USEFOR WG
has taken the position that the only one worthy of any official
recognition at all is a single "Re: ", and that even that usage might be
better it it were phased out (some hope!).

Within the context of a mailing list (essentially a closed community),
there _might_ be some justification for the convention discussed in this
draft, but it still leads to many complications, not least the fact that a
message cross-posted to two mailing lists will have a different Subject in
each list, which is going to lead to threading problems if you manage to
get followups also crossposted to both lists (a desirable aim, but very
difficult to achieve, as discussed here recently in other threads).

There are a lot of problems with mailing list (again see recent threads
here) and essentially what is needed is a BCP document (perhaps with a
standard-track companion if it turns out new headers such as MFT are also
needed) to establish just what mailing list practice should be. The
question of these header tags could well have a place in such a document.

Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web:
Email: chl(_at_)clerew(_dot_)man(_dot_)ac(_dot_)uk      Snail: 5 Clerewood Ave, 
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5

<Prev in Thread] Current Thread [Next in Thread>