> And, as Jon Postel once said, "this is a naming system, not a general > directory assistance system". So, like Ron explicated, there is > absolutely no need to behave under a domain in such a way as to respect > necessary and sufficient conditions for semantic equivalence (or > non-contradiction) with the stated meaning of the string. Jon Postel's obvious wisdom notwithstanding, by the time ICANN was created the view of most was that the DNS for Web URLs was exactly such a directory service. So many many problems we now have are explainedd by that view becoming dominant. It should also be noted, though, that when Dr Postel made his statement, the Web wasn't in existence and Hypertext generally was in its infancy. > In any case, under the financial conditions of new gTLD applications, > gunning for a .peadophile TLD might not be all that rationnal. I'm > guessing that the high price tag on new gTLD application is protection > enough for obviously inesthetical or frivolous TLD registration (am I > already contradicted by experience?). Oh, I agree completely that it's highly unliklely that under the current funding proposals .paedophile is unlikely to be suggested. However, it used to cost $200 pa (when $200 was worth more, as well) to register a domain name, and within a couple of years it came down to $20. Basing our acquiescence to a framework on the initially prohibitive cost (which we've already complained about in respect of, for example, developing economies) would possibly leave us with egg on our faces should the price quickly drop to $10,000 or $1,000 in a few years. > Obviously, i can see that we could want to "give-a-little" on such > issues to the GAC if they are sleepless about it and, in turn, are > menacing some other area of import to us (i have no example, and am > merely speaking in the abstract). Conversely, they must always be on the > lookout to expand their reach, so from this perspective, then we are > obviously at odds with them ... Whether and how much to compromise is always part of real-politik in such situations, but often in ICANN we seem to find ourselves with a strong view which ICANN doesn't share and so the question them comes as to whether sticking to the strong case gets us no say in what actually happens or whether it means the end result is slightly less bad. The danger of the current proposal from staff is that we're not just giving a little, but creating something at least as bad as the worst excesses of UDRP and possibly much worse. The compromise suggested at the start of this discussion in NCSG/NCUC is already giving a little, and we have the joint support of ALAC for the stance of giving a little to avoid giving away most if not all of the field. > Andrew, how does internationalisation mix in? phishing attempt by way > of, say, cyrillic caracaters? Then i guess i would support objections > based on "confusing similarity" with another tld. This is one of the issues about internationalisation, certainly. THere are potentially many others. Roman representations for systems without the character set comes to mind. There was a (in)famous example of a two-letter encoding of Veitnamese used before Unicode was common where two of the most common phonemes in Vietnamese were represented as "IT" and "SH", leading to all sorts of fun with GENie (IIRC) English bad-language filters. There's the "scunthorpe" issue multiplied into other languages (take letters 2-5 of that UK place name if you don't know this one: the French equivalent word uses an "o" instead of a "u" and the ISP in question allowed "sconthorpe" but not "scunthorpe" to be registered as a userid). I don't know of any, but there may well be Japanese/Chinese kanji/hanzi words where it's normal in one language but rude and offensive in the other. -- Professor Andrew A Adams [log in to unmask] Professor at Graduate School of Business Administration, and Deputy Director of the Centre for Business Information Ethics Meiji University, Tokyo, Japan http://www.a-cubed.info/