Re: IDN problem.... :(

David Clarke wrote on 2/15/2005, 10:06 AM:

 >
 > One possible solution for the IDN is to determine "similar" looking
 > characters, such as I (U+ 0049: Latin Capital Letter L)  and l (U+006C:
 > Latin Small Letter L) (which look visually identical in Arial Unicode
 > MS) and limit new  registration of only one of the code points.

The differences between 'I' and 'l' usually can solve by case folding to 
lower case characters. Once 'I' is case fold to lower case, it show 
differently in the URL bar as 'i'. The problem is between 'l' and '1' 
(one).


 >
 > This would also cut the cost of registration fees for the companies.
 >
 > This of course does not totally solve the problem, but it would limit
 > the spoofing.
 >
 > Areas that it would not solve the problem would include Japanese Kana
 > some of which have small and large forms, which are use differently but
 > often need to be mixed in a word or name. It is conceivable that a
 > company whose name is entirely "big" Hiragana could be spoofed with one
 > made up of small Hiragana only, giving very little difference in
 > appearance, apart from size.

I think those issue was solved in IDN already in the name prep.

 >
 > Note also that in the case of Hiragana or Katakana used this way, the
 > "colouring" solution would not distinguish, as they are both part of the
 > same Unicode sub range. In fact the small and large versions of the
 > characters are consecutive pairs of characters.

Again. forget about the coloring approach. Remember your color blind 
neighbor (as well as W3C accessiblity guideline), and some future 
devices which may be black and white in their first version.

 >
 > --
 > David Clarke
 >

Received on Wednesday, 16 February 2005 16:11:00 UTC