Re: SIMPLE chat: Internationalization and normalization of nicknames
Peter Saint-Andre <[email protected]>
| Newsgroups | gmane.ietf.simple |
|---|---|
| Message-ID | <[email protected]> |
On 3/2/12 12:08 PM, Ben Campbell wrote: > > On Mar 2, 2012, at 10:39 AM, Paul Kyzivat wrote: > >> On 3/1/12 5:54 PM, Ben Campbell wrote: >> >>>> OK. It seems that you might also be making some authorization >>>> decisions based on the nickname, e.g., if you've reserved a >>>> nickname then I can't get that nickname (can I join the >>>> chatroom but just not with that nickname?). >>>> >>> >>> Yes, that is true. We do need an equality operator in some sense. >>> (And yes, assuming no other policy issue was blocking you, you >>> could still use a different nickname) >> >> And this presumably needs some notion of canonicalization, so that >> you can't get two nicknames that compare as different but *look* >> the same. > > That's going to be a very hard problem to solve. I suspect we will > end up with text that points out that this can happen, but leave it > up to implementors to figure out what to do about it. The problem of so-called confusable characters is impossible to solve. See for instance: http://tools.ietf.org/html/draft-ietf-precis-framework-01#section-10.3 The Unicode Consortium has gone a bit farther: http://unicode.org/reports/tr39/#Confusable_Detection You can nibble around the edges here, but foolproof solutions cannot be had (in large measure because humans are so easily fooled). > (Peter: Do you have your Cherokee example handy?) http://stpeter.im/journal/1420.html That is: ᏚᎢᎵᏋᎢᏋᏒ != STPETER (and Cherokee Letter Quv = Ꮛ is not defined by the Unicode standard as confusable with, mapped to, or otherwise equivalent to Latin Uppercase E despite their surface similarity). Peter -- Peter Saint-Andre https://stpeter.im/ _______________________________________________ Simple mailing list [email protected] https://www.ietf.org/mailman/listinfo/simple