Re: Cross-WG WGLC of draft-ietf-xmpp-6122bis-18

Hi Dave, thanks for the review. Comments inline.

On 3/5/15 9:57 AM, Dave Cridland wrote:
> On 27 February 2015 at 02:17, Ben Campbell <[email protected]
> <mailto:[email protected]>> wrote:
>
>     We actually did a WGLC on this about a year ago, but much has
>     changed since then. Please send comments to the authors and both
>     lists by March 12, 2015. If you've read the draft and consider it
>     ready to go, please send comments to that effect.
>
>
> Non-Precis comments:
>
> 1) "at slash at".
>
> The domainpart of a jid is described as being after the first '@', if
> any. That's not always true; consider the following legal jid:
>
> service.example.org/[email protected]
> <http://service.example.org/[email protected]>
>
> This issue arises because the resourcepart can contain an '@', and to
> find the domainpart we need to (conceptually) remove the resourcepart first.

True.

> So I think this implies two things: Firstly, the text at the beginning
> of §3.2 needs changing slightly, perhaps:
>
> The domainpart of a jid is that portion which remains once any portion
> from the first '/' character to the end of the string has been removed
> (if there is a '/' character present), and then any portion from the
> beginning of the string to the first '@' character (if there is a '@'
> character present).
>
> This is ugly text.

Well, it is accurate, which is an improvement.

> The second implication is that the examples could use the nasty cases
> like this.

Yes. Will add.

> 2) "case folding examples"
>
> The examples given as 8, 9, and 10 confuse me more by their presence
> than their absence, because the resourcepart of a jid is compared
> case-insensitively anyway - I don't know Greek well enough to know if
> the final sigma would be used at the end of a localpart, but it seems to
> me that the capital and lower case sigma example might be best re-cast
> as a localpart example.

That's a good suggestion, thanks.

> (When I read the details of the case-map versus final examples, I
> worried that we'd changed the comparison for resourceparts to
> case-folding somewhere...)
>
> 3) "angle brackets"
>
> The presence of angle brackets in example 16 is, I think, more confusing
> than the thing it's attempting to solve; in part because the < does
> indeed make the jid illegal under §3.3.1, so I didn't initially read the
> explanation...

I think it would be best for all of the examples to be enclosed in angle 
brackets (this is what we did in the precis-nickname spec).

> 4) "applicability to what now?"
>
> Can we change §3.4.1's title to "Applicability within XMPP extensions"
> or something?

Sure.

> I'll see if I can navigate the Turtles of Precis enough to come up with
> some comments there. Sadly I suspect I won't be able to sensibly comment
> until I've tried implementing.

That's quite possible. Internationalization is perplexing.

Peter

-- 
Peter Saint-Andre
https://andyet.com/