Re: My prod at IDN requirements

["Martin J. Duerst" <duerst@w3.org>]

At 09:44 00/01/05 +0100, Harald Tveit Alvestrand wrote:
> At 11:11 05.01.00 +0900, Martin J. Duerst wrote:
> 
> >Why not? Couldn't somebody create something like WWW.1BM.c0m, and try
> >to confuse people?
> 
> Already done. See micros0ft.com (currently located offshore somewhere, 
> after exchanging angry letters with Microsoft advocates).

Interesting example. At the moment, it does not look that bad
(they have a disclaimer at the bottom, with a link to Microsoft),
but then I don't know how it looked previously.

Also, it seems that because DNS names are preferably written
lower-case, the chance of the reader of a domain name/URI getting
confused is rather low. This would extend to Greek and Cyrillic;
the similar letters among these scripts are significantly fewer
in lower case than in upper case:
Greek: o
Cyrillic: a, e, o, c, y, x (+ s, i, j, and a few with diacritics)
[Historical note: This is due to the fact that the scripts split
off when there were only upper case letters; the three scripts
later took different ways to create lower case letters. In
Latin, the lower case letters used in print have been fairly
stylized back from their cursive origin, in Greek, they stay
rather cursive even in print, and Cyrillic is mostly using
smaller copies of the upper-case letters for lower case in
print while there is a separate cursive tradition for handwriting.]

If we stay with this analysis, then it looks to me as if the
danger for such cases is not really increased when going from
ASCII to something else, and so I don't think we should need
to be more strict than for ASCII.

Also, I wonder whether for the above example, there is a big
difference between micros0ft.com and e.g. microsuft.com. The
later is probably even more difficult to detect. And while
one can think about how to rule out the former, simple rules
to eliminate the later don't seem possible; they would be
highly language and culture-dependent.


Anyway, Harald, as you know this example, maybe you know others
that may give us more insights.


Regards,   Martin.


#-#-#  Martin J. Du"rst, World Wide Web Consortium
#-#-#  mailto:duerst@w3.org   http://www.w3.org