From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Mon Jul 01 2002 - 05:47:01 CDT
In <3D1CAD61.8040508@certplus.com> Jean-Marc Desperrier <jean-marc.desperrier@certplus.com> writes:
>Charles Lindsey wrote:
>>But OTOH we don't want to give it even the respectability that a "SHOULD
>>NOT" might imply, so I carefully avoided that by say explicitly that it
>>"is not compliant with this standard". That way we give the right message,
>>whilst not opening ourselves to yet another accusation that "USEFOR is
>>riding roughshod over all existing practice...".
>>
>I can't agree with that.
But I don't see exactly what your disagreement is.
>The way to accomodate with existing with existing practice is given
>below when explaining what to do when recognizing that headers are not
>in UTF-8.
Yes, it says you MAY do that. Someone has asked for words such as "best
current practice" to be added.
>The standard must be clear that the continuing use of unencoded local
>encoding is not conformant.
Yes. That is exactly what my text says.
>It may not make everybody happy, but the reason for that will be that
>they want to change nothing, and if we go that way, we will stay with a
>non-working situation forever.
Yes, that is what we want to achieve. But you can't quite say MUST NOT,
because there is too much existing practice. OTOH if you say SHOULD NOT,
then they will say "Ah! I see it is not absolutely forbidden, so I will
carry on doing it". So I wrote that it was "non-compliant" in the hope it
would give the right message.
>Most newsreaders understand RFC2047 in headers.
>When Agent 1.92 comes out, only outdated or very exotic users agent will
>be unenable to decode them.
There are still a lot of people using trn and the like. Yes, as a
percentage of Usenet they are declining, but that is because of the flood
of newbies using newer stuff, not because any old hands are giving up
what they have used these past 10 years :-) .
>> Some reading agents do not recogize
>> [RFC 2047], and some are incapable of decoding UTF-8
>This is not true
Yes it is.
>Some agents will just transmit everything directly to the output device,
> if it's configured to UTF-8, it will go through, but this is not a
>usable setting in a world as we know the most common usage is other local
>encoding, but *not* utf-8.
Yes, but I include the underlying OS under the term "reading agent". Does
OE understand UTF-8 when running under Windows 95/98?
>The note ought make it clear that for interoperablity, RFC2047 will be
>better than raw UTF-8.
Yes, but other people on this list are saying that they want UTF-8 to be
the ultimate method of choice. I cannot satify both parties, so I have
left it without any strong recommendation (note that there is a sentence
enclosed in {...} suggesting UTF-8 as the ultimate goal, which Russ has
asked to come out and I think you would agree - I am just waiting for
more opinions on it).
-- 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@clw.cs.man.ac.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