From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Tue Mar 04 2003 - 07:54:59 CST
In <200303030547.h235l910019186@lima.epix.net> "Forrest J. Cavalier III" <mibsoft@epix.net> writes:
>> No, becausde if a future draft wanted to do I18N names properly, using
>> UTF-8, it would not be compatile with an interim I18N of names using
>> punycode.
>1. How will an existing punycoded name interfere with creating a raw UTF-8
> newsgroup name (_if_ they were to be standardized?)
Because it will bring about a state of affairs where some groups in some
hierarchies will have a canonical name in punycode and some other groups
in some other (or even the same) hierarchies will have a canonical name in
UTF-8. Which will be the cause of the most utter chaos, with posting
agents not knowing which to generate (since what the user sees or
types will appear to be ordinary characters in a, to him, plausible
charset).
Or it will lead to various agents erroneously trying to convert one to the
other, and active files that contain both forms, and goodness knows what
else.
>2. Do you understand the inconsistency that you think it is fine that
> the present draft requires upgraded injection agents everywhere (to
> allow moderation to work), but you think it impossible that a
> future draft requires upgraded injection agents (to convert punycoded
> names to raw UTF-8)?
Please read what is proposed in my proposal entitled
I18N of newsgroup-names using UTF-8
(which is more restrictive than the present draft). It does NOT require
upgraded injection agents everywhere from day 1 (i.e. it reduces a 'Y' in
Russ's scheme to a 'D'). It does this by saying that moderated I18N groups
Ought Not to be created until a sufficient base of upgraded injecting
agents is in place (you can make that SHOULD NOT if you like, but our
convention has been "Ought Not" when giving advice to hierarchy
administrators).
>3. Since the IETF is in charge, there will be no USEFOR draft of I18N names
> which requires upgrading injection agents. So the previous points are moot.
That remains to be seen. All we have had so far is threats from the Area
Director (the chief effect of which was to force our Chairman to quit).
When we have a new Chairman in place we will be in a position to
negotiate with the full IESG. It is by no means clear that UTF-8 will fail
for newsgroup-names even if it fails for other stuff. It will depend on
whether we can convince them that the technical details are mail-safe.
And, as I have said, I18N newsgroup-names does not "require upgrading
injection agents". It is only the moderated ones that do that.
>4. punycode does implement I18N "properly", but it may not have your "blessing."
> Don't confuse religion with engineering.`
Confusion that may result from the future coexistence of punycode and
UTF-8 newsgroup-names is most certainly an "engineering" issue.
-- 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