I agree.<br><br>Mark<br><br><div><span class="gmail_quote">On 12/21/06, <b class="gmail_sendername">John C Klensin</b> <<a href="mailto:klensin@jck.com">klensin@jck.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br><br>--On Thursday, 21 December, 2006 19:08 +0900 Martin Duerst<br><<a href="mailto:duerst@it.aoyama.ac.jp">duerst@it.aoyama.ac.jp</a>> wrote:<br><br>> At 05:44 06/12/21, John C Klensin wrote:<br>><br>> [removing most of the post, because I generally agree]
<br>><br>>> There is another problem with prohibiting script-mixing at the<br>>> protocol (IDNA) level and that is that the common,<br>>> on-the-street, perception of "the script we use" is different
<br>>> from the Unicode definitions of "script". No one is wrong<br>>> here, but, if JDNC concludes that Romanji is a necessity and<br>>> must be available in mixed names with Kanji and Kana, I don't
<br>>> think we are in a position to say "no" (although we can<br>>> _advise_ that this isn't a good idea).<br>><br>> And why _should_ we advice that it isn't a good idea?<br>> The confusion potential between Latin and Kanji/Kana is
<br>> virtually nil.<br><br>I was not proposing that we do so, merely trying to identify the<br>contrast.<br><br>>> Similar examples arise with mixtures<br>>> of Cyrillic and Roman characters in Russia, even though we are
<br>>> agreed that is one of the more dangerous cases of mixed-script<br>>> labels (the fact that some strings in Cyrillic can be confused<br>>> with names in Latin characters even when they are purely<br>
>> Cyrillic is one of the arguments why prohibiting mixed scripts<br>>> isn't nearly as powerful a tool as is often argued).<br>><br>> Yes. The amount of danger comming from script mixtures depends<br>
> extremely strongly on the scripts involved. That's why any kind<br>> of general solution, even in the form of a recommendation,<br>> is probably a bad idea.<br><br>We are in complete agreement, I think. The only recommendation
<br>I would recommend (sic) would be that registries study the<br>scripts that they permit to be mixed very carefully and make<br>policies that they believe reflect an appropriate balance for<br>their user and registrant populations. I think we might
<br>rationally couple that advice with a warning that characters<br>that might look very different to experienced users of the<br>scripts involved might look similar enough to be confused by<br>those who have less experience.
<br><br>We do need to keep in mind that advice of that sort is likely to<br>be almost useless to most operators of TLDs with global scope.<br>Their problems with the implicit requirement of treating all<br>languages and scripts equitably is, as always, much more severe
<br>than the problems faced by a ccTLD that can decide on support<br>for a relatively smaller number of scripts or writing systems.<br><br>That conclusion has two corollaries that I think we need to<br>understand. The first is that trying to build a "no mixed
<br>scripts in a label" rule into the protocol is probably an idea<br>that isn't going anywhere (and shouldn't). The second is that<br>any rule in application software that treats mixed-script labels<br>as inherently dangerous should be extremely localized, probably
<br>examining the user's preferred languages or scripts and the<br>particular combinations of scripts being mixed, lest one<br>generate warnings about many safe situations and thereby cause<br>users to discount warning information that really is important.
<br><br> john<br><br></blockquote></div><br>