Re: URL internationalization!

Yergeau;

> This one is particularly ironic after you implicitly recommend what is no
> more than a dirty hack to solve a long-standing interoperability problem,

No. It's Martin who has been persistingly trying to introduce the
interoperability probelm.

> instead of fixing the spec that creates it. The meaning of non-ASCII octe=
> ts
> is unspecified by RFC 1738 and its successor-to-be (your draft), and the
> best you can come up with is to solve that by advising each and every for=
> m
> author to place a marker in each form so that the server can *guess* the
> encoding?

No. The ovbious answer is for the client to use ISO 2022 based
encoding assuming G0 designation of ASCII only.

The encoding is compatible (automatically distinguishable with 100%
certainity) with all the existing ones, including SJIS, EUC (of any
sort including ISO-8859-*), BIG5, GBK, and even the infamous UTFs
so that servers can automatically detect badly-behaving clients
and, then, guess the encoding.

							Masataka Ohta