In <200801211855(_dot_)34754(_dot_)Mark(_dot_)Martinec+ietf(_at_)ijs(_dot_)si>
Mark Martinec <Mark(_dot_)Martinec+ietf(_at_)ijs(_dot_)si> writes:
Charles Lindsey wrote:
I think it is clear (3 people against it) that the wording would have to
avoid "SHOULD". So here is a suggested wording avoiding 2119 language:
Although in all header fields the CFWS following the colon is optional, it
is customary practice to place at least one SP there; moreover such a SP
is mandatory in Netnews [UESFOR] (and possibly in other protocols).
Continuance of that practice by implementors is encouraged in order to
facilitate interoperability.
That could probably be tuned further.
The text looks fine, but I would like to point out one catch
which could result from placing too much emphasis on that space.
Namely, a header field consists of a header field name, a colon,
and a header field body, i.e. everything after a colon belongs
to a field body, including that [CFWS]. It would be wrong if
it could somehow be undestood that a field body starts only
*after* a space.
I think the "man in the street" would be surprised to be told that initial
(or even trailing) whitespace was a part of the field-body (though he
would assuredly regard any <comment> as being part of it).
So first question, is whether that is really the best definition of "field
body"?
And the second question is what differences would arise if it were defined
differently?
I think we need to understaand the implications of question 2 before
deciding whether some further tuning of the text is needed.
--
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_)clerew(_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