On 1/5/06, Tom.Petch <sisyphus@xxxxxxxxxxxxxx> wrote: > You say that a Unicode code point can be represented by %xABCD but that is not > spelt out in ABNF [RFC4234]. ABNF uses non-negative integers to represent characters - note that it explicitly doesn't specify a range (2nd sentence of section 2.3). RFC 4234 specifies the encoding of those integers into ASCII, and says that other encodings may be specified. There's an implicit (obvious) encoding of ABNF characters into Unicode, and as Misha pointed out the IRI spec uses it - technically maybe it needs to be spelled out somewhere. > And when it refers to 'one printable character' as > '%x20-7E' I get the impressions that coded character sets like Unicode, with > more than 256 code points, do not fall within its remit. That's an example, which is probably missing the word "US-ASCII" between "printable" and "character". Bill _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf