--On Monday, July 20, 2015 1:57 PM -0400 Ted Lemon <ted.lemon@xxxxxxxxxxx> wrote: > On Jul 20, 2015, at 1:15 PM, John C Klensin > <john-ietf@xxxxxxx> wrote: >> I agree with your analysis, modulo one question. You make the >> comment that CLASSes don't really work, as have others. I'd >> like to understand it better. > > To understand why classes don't work, can you tell me what a > URL for a name in a non-IN class would look like? Depends on your perspective. If we had used CLASSes for IDNs (a bad idea for other reasons, but seemingly interesting at the time), the answer would have been "detect non-ASCII characters in the domain". More generally and with the understanding that it doesn't actually help us make progress, it would be plausible to answer your question above with "CLASSes work fine, it is URLs that are broken and don't work". Seriously. One would need some lexical mechanism to determine whether an apparent domain in a URL is to be resolved in CLASS=IN or some other CLASS. That is no such mechanism. but, while the makes CLASSes hard to deploy, the fault lies with the RUL/ URI system and its having become some complex and rigid that it is difficult or impossible to extend and adapt to previously-unidentified requirements. john