Re: Fwd: Last Call: 'Atom Threading Extensions' to Proposed Standard (draft-snell-atompub-feed-thread)

[Ned Freed <[e-mail]>]

> FYI -- this is an individual submission yet intended for Proposed
> Standard, so input is most welcome.

I have a couple of comments, all minor.

The sentence at the beginning of section 3 is confusing (to me at least). it
says:

  The "in-reply-to" element is used to indicate that an entry is a
  response to another resource and MUST contain either or both the
  "ref" and "href" attributes.

Unless you read it carefully it sounds like the MUST requirement applies
to the referenced entry, not the in-reply-to element itself. How about
changing it to something like:

  The "in-reply-to" element is used to indicate that an entry is a
  response to another resource. The element MUST contain either or both the
  "ref" and "href" attributes.

Now, I have a lot more experience with XML Schema than with Relax NG so 
maybe I'm reading this wrong, but the Relax NG definition seems to say
that the source attribute can only appear when the ref attribute is present
and the type attribute can only appear when the href attribute is present.
But this isn't spelled out in the (normative) text. How about adding
"When the ref attribute is present" before "The 'source' attribute ..." and
"When the href attribute is present" before "The 'type' attribute ..."?

Finally, the document seems a bit short on specifics of how the 
in-reply-to element is actually used. Although the underlying semantic
model here seems to be simpler than email's in-reply-to/references scheme
(a good thing IMO), perhaps some words about whether you need to list just
the parent(s) and not the grandparent(s) would be in order. There
were certainly differences of opinion about this in the days of RFC 822
which RFC 2822 section 3.6.4 cleared up.

That's it!

				Ned