[idn] Editorial comments on punycode

[Erik Nordmark <Erik.Nordmark@sun.com>]

draft-ietf-idn-punycode-01.txt
------------------------------

I've reviewed the punycode document to make sure it is
ready for IETF last call.

Here are some editorial comments that it would be useful to
fix before the last call.

Overall the document looks very good.

The introduction of "bootstring" in the abstract is a bit odd - almost
made me look for bootstring in the references section.
Saying "This document defines bootstring as a general algorithm ..."
is more clear.

Section 4 talks about mixed-case annotations but it hasn't been mentioned
earlier in the document - a bit surprising to the reader. Thus it would be 
good to introduce this optional possibility earlier in the
document. E.g. in section 1.2 stating that this optional capability
exists in the algorithm but that it isn't needed for IDNA.

Section 6.2 says at the end what the decoder need not re-encode its output.
Why would it ever need that? I think there is some context missing but I
can't guess what that context is.

Section 6.3 talks about overflow - but is the information in that paragraph
common to the decoder and encoder? (The text lives only in the encoder 
section.) If not what are the size requirements for the integer data type
for the decoder?

Section 7.1 talks about "the IDNA signature prefix" but the IDNA
spec uses the term "ACE prefix".

Split normative vs. non-normative references.

---