In <3E1C298F(_dot_)4090502(_at_)Sonietta(_dot_)blilly(_dot_)com> Bruce Lilly
<blilly(_at_)erols(_dot_)com> writes:
Now if I had
Return-Path: <(?-?-?-?)>
where "?-?-?-?" should be regarded as a placeholder for some sensible
comment in Hungarian, Arabic or Chinese, then my proposed workaround
would transform it into
Return-Path: =?iso-8859-1?q?<(=3F-=3F-=3F-=3F>)?=
which, I will grant you, is not a valid Return-Path
As you may recall, that was in response to the lastest variant of your
gateway "heuristic", in which you claimed that URIs should be bracketed
by '<' and '>' on the supposed grounds that that would prevent
misinterpreting parentheses in URIs as comments, further supposing
(incorrectly) that comments could not appear within angle bracketed
constructs. So your gateway process (which you now call a "workaround")
would in fact entirely miss the content within the angle brackets. And
that proposed gateway heuristic simply won't work precisely because
it does not take into account full syntax.
Exactly so. A "heuristic" is, by definition, a method that will get the
right answer in a sufficiently high proportion of cases to be useful.
Inevitably, it will fail in a few cases. The art of designing a heuristic
lies in knowing which cases are likely to happen in practice, so that the
benefit can be maximised. The occurrence in a Netnews article of a
Return-Path containing non-ASCII within a comment is as unlikely as they
come.
--
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