At 01:52 05/01/26, Paul Hoffman wrote:
>
>At 9:16 AM +0100 1/25/05, Julian Reschke wrote:
>>I saw some concerns (with which I agree) that requiring the presence of an IDN library is problematic. As far as I can tell, it's likely to be ignored by developers of clients that run on somwehat constrained devices.
Just a correction to what Julian wrote: Stringprep may be ignored or partially ignored, because it requires lots of data, but I don't think the Unicode->punycode conversion will be ignored.
>I would like to hear more from developers whether or not they think this is a problem. (I'm asking this wearing my co-chair-looking-for-consensus hat, not my author-of-IDN-and-stringprep hat.)
Although I don't have any experience with developing blogging tools, I have some experience in integrating IDN support into a browser, and I can report on some other browsers. I hope this breaks the ice for others to chip in.
1) I have integrated IDN support into libwww, and used it in Amaya (that version isn't yet public, but it should be soon). Integrating idnkit into libwww was easy (except for the autoconf/ automake,... files, which I'm leaving to somebody else). idnkit of course includes full stringprep.
2) The maker of a widely used browser for Mobile phones has just recently announced that they support for internationalized domain names at http://www.access.co.jp/english/press/050120.html My guess is that they take some shortcuts on stringprep, but that's really just a guess. If you really need to know, I may be able to ask somebody from that company.
3) Browsers start to integrate blog support. These browsers are at the leading edge of development, and usually support IDNs already. The examples I know are Opera and Safari. IDNs will just work on those browsers, and people will come to expect them to work in other contexts. I think at least at some stage, Opera didn't implement strinprep (or not completely), but I haven't tested lately.
Regards, Martin.
