In <p0510030fb8888960ce0a(_at_)[130(_dot_)237(_dot_)150(_dot_)141]> Jacob Palme
<jpalme(_at_)dsv(_dot_)su(_dot_)se> writes:
At 10.49 +0000 02-02-07, Charles Lindsey wrote:
So you need a requirement that things only go into the registry if it is
anticipated that widely deployed software will want to parse and
understand them (and which it would be a pain to rewrite if it had to be
changed from X- to non-X- later on).
That is too strict. If an experimental system wants to try
out a new facility, like "Counter-Argument-To", this should
be allowed even if we do not know if it will be widely
used. The importance with the registry is (a) that it stops
someone else from using the same header name with a
different usage (b) someone who wants to implement a new
header can find out and learn from, and make its
implementation compatible, with what others have done (c)
it can warn against dangerous header names. None of these
advantages with a registry will work if we only register
widely deployed headers.
No, you missed my use of the word "anticipated".
If 3 or 4 sites want to use Counter-Argument-To amongst thjemselves for a
very special purpose, then it should be X-Counter-Argument-To.
If 3 or 4 sites say "let us try this out strictly between ourselves", then
it should be X-Counter-Argument-To.
But if they say "this seems a neat idea; let us try it out to see if it
works and if it does then maybe it will catch on worldwide", than they have
"anticipated" a wide deployment, and they should use Counter-Argument-To,
having first got it into our proposed register (which implies passing our
modest hurdles, of course).
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: chl(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk Snail: 5
Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5