Re: some comments about draft-ietf-idn-idna-08.txt

Paul Hoffman / IMC <[email protected]> wrote:

> it looks bad if one of the authors says "we need to change this"
> without all of us having thought about it.

You're right, sorry.  Feel free to override the Reply-To field on this
message.

> Why do you want the change?

The big problem is that this sentence got lost between drafts 7 and 8:

    ToASCII fails if any step of it fails.

The new text doesn't say anything like that, but it's a crucial part of
the definition of ToASCII.  That's why the commentor was confused about
what happens if step 3, 5, or 8 fails.

A much much less important issue is the choice between these:

    If any step fails, the original sequence MUST NOT be used as a label
    in an IDN.

    If the ToASCII operation fails on any label in a domain name, that
    domain name MUST NOT be used as an internationalized domain name.

I think they both amount to the same thing.  Neither one suggests an
alternative, and they both imply that the name you're working on is
no good.  I prefer the former wording because ToASCII operates on an
individual label, oblivious of the other labels in the same name, so
it's nice if the text in the ToASCII section likewise makes no mention
of the other labels in the same name.  The latter wording still appears
in section 1.1.

If you're not convinced about this sentence, it's no big deal, but the
other sentence is important.

AMC