RE: [idn] RE: An idn protocolfor consideration in making the requirements

[Karlsson Kent - keka <keka@im.se>]

Title: RE: [idn] RE: An idn protocolfor consideration in making the req uirements

 

-----Original Message-----
From: Jonathan Rosenne [mailto:rosenne@qsm.co.il]
 ... 

For a fair number of people in the world, ASCII is barely readable, and their scripts are unreadable many others. An I18N design, while it has to accommodate to the existing systems in place, should should be script neutral and not biased in favor of English.

 

If I understand your proposal correctly, in essence you propose that the official name will be ASCII, with synonyms in other scripts. The non-ASCII names will only be used if the user agent in some way asks for them.

The idea is that the *fallback* name would be in ASCII (for SMTP mainly; HTTP apparently don't need this even now!).  In many cases the *preferred* name(s) would  not be in ASCII.  If that happens to (in some way) imply that the "official" name would be in ASCII, does not worry me, as long as it is only used as a fallback when there is another (preferred) name or names.  I'd rather have a fallback to something barely readable than having something which is definitely unreadable.

I think the problem you raise is not a real problem, it is just a nuisance. People will not very often be exposed to "gibberish" ASCII names, 

 

The experience with BASE64 for text and QP strongly indicate otherwise.  Even after 10 years not even IETF nor IMC themselves can handle it properly (see ftp://ops.ietf.org/pub/lists/idn.current [which is 'in the raw'] and http://www.imc.org/idn/mail-archive/ [where BASE64 is not decoded in headings and QP looses its charset declaration]).  And the number of bugs, and the time to (partially!) correct them, relating to QP and BASE64 in e-mail products is not encouraging for that kind of approach.

 

 they will have to ask for them in some way or receive mail. People who will be intentionally using IDNS will necessarily have a suitable user agent. And if you are exposed to these names in the raw, they are just meaningless, not causing any other difficulty. If you reply to such mail, it should still work.

Nobody in the right cautious mind would use them for reply.  They would, however, cause questions to someone that "might know" on why the h--l the address suddently (and sometimes) turned into crap.  There would be no easy and good answer.  And there would be no end to that.

Jony 

 

        Kind regards

        /kent k