[standards-jig] JIDs (JEP-0029)

[Craig]

Replies inline:

>1. I would prefer to see consistency between this JEP and the relevant
>subsection of the IETF draft regarding terminology. For example, there is
>a reason I started to use the term "node" as opposed to "user" in the IETF
>draft, since the portion of the JID before the at sign is not always a
>user, but rather may be something like a conference room (in groupchat) or
>a publisher (in some pub/sub-style applications). Such entities are not
>users. 
>

Agreed.  I'll use your nomenclature.

>2. I would prefer that this JEP make it clear that only the domain
>identifier is necessary to have a valid JID. The node identifier and
>resource identifier are optional.
>
See the first line of the grammar:

<JID> ::= [<user>"@"]<host>["/"<resource>]

The brackets around user@ and /resource mean that they are optional.

>
>3. Characters disallowed in usernames (node identifiers, whatever :) are
>specified as Unicode character numbers in the IETF draft (e.g., U+003E)
>but by ostensive example in the JEP (e.g., >). I would prefer that we not
>resort to ostensive definition.
>
Actually, I specified allowed characters in the grammar:

<conforming-char> ::= #x21 | [#x23-#x26] | [#x28-#x2B] | 
                            [#x2D-#x39] | [#x3B-#x3F] | 
                            [#x41-#x7E] | [#x80-#xD7FF] | 
                            [#xE000-#xFFFD] | [#x10000-#x10FFFF]


The disallowed characters were only spelled out for those who don't know 
what characters I excluded in this terminal.  I can wipe out the 
offending list from section 2.3 since it adds no information not already 
specified in the grammar.

>
>4. I'm not a character encoding guru -- is there a preferred style of
>identifying Unicode characters (e.g., U+0020 vs. #x20)?
>
Well, I went with the same notation used in the XML spec 
(http://www.w3.org/TR/REC-xml).  Seemed good enough for me ;-)

--C