AMC wrote: > > Regardless of what forms are possible/forbidden in DNS responses, people > will still be able to present the names in whatever equivalent forms > they like, for better or worse, to trick people or to make the name more > readable. I agree. WWW.LOOT.COM takes you to an eBay wannabe buying/selling site. WWW.L00T.COM is also a registered domain, and goes to an EasySpace client. Coincidence? Probably not. This kind of stuff is done all the time, and it doesn't depend on glyph lookalikes to create confusion. Similar spellings, semantic replacements, or substitution of plausible TLD's can all do the job. I don't see how moving to Unicode for IDN's makes the problem significantly different, except that it exposes DNS implementers to a much wider variety of characters that could be visually confused for each other. I don't think the IDN WG concerns with nameprep should extend beyond ensuring that the output of nameprep doesn't produce any characters that interfere with the formally reserved characters for URL syntax and the like. Beyond that, just let the Unicode equivalence tables take care of things. --Ken > > AMC > >
