Re: [xmppwg] Re: Last Call: 'XMPP Core' to Proposed Standard

[Peter Saint-Andre <stpeter@jabber.org>]

Comments and proposed text at end.

On Sat, Sep 27, 2003 at 09:39:41AM -0700, Pete Resnick wrote:
> On 9/26/03 at 12:24 PM -0500, Peter Saint-Andre wrote:
> 
> >   Therefore I propose that the following text be added to
> >   Section 4.2 (Stream Attributes) of this memo:
> >
> >      o  xml:lang -- An 'xml:lang' attribute (as defined in
> >      Section 2.12 of the XML specification [1]) SHOULD be
> >      included to specify the default language of any XML
> >      character data sent over the stream.  The language
> >      specified for the stream MAY be overridden by the
> >      'xml:lang' attribute of a specific XML stanza or a
> >      child element thereof (see xml:lang (Section 8.1.5)).
> >      The value of the attribute MUST be an NMTOKEN and
> >      MUST conform to the format defined in RFC 3066 [15]. 
> >      If no 'xml:lang' value is provided for the stream,
> >      an implementation MUST assume that the language is
> >      "en-US".
> 
> I had a conversation with Ted and now I see what his concern is. Let 
> me explain what I think should be done. There are two parts:
> 
> 1. The last sentence of the above is simply wrong. The idea that 
> there should be an "English default" was introduced when we were 
> talking about 'xml:lang' as some sort of locale request, and we came 
> to consensus that we didn't want that. The only thing 'xml:lang' is 
> telling you is the language of the contained text. Given that, we 
> don't want to say that it defaults to English: What would it mean to 
> say that some Unihan character should be interpreted as English? So, 
> strike the last sentence.
> 
> 2. Ted is concerned that since the XML spec says that the 'xml:lang' 
> applies to *all* attribute values, implementors of future protocols 
> based on XMPP (i.e., something other than IM) need to be warned that 
> even though 'xml:lang' is on the entire stream, it only applies to 
> text intended for human consumption. (His fear is that someone might 
> do some goofy language dependent processing on an attribute value 
> that is clearly not human readable text.) I think this can simply be 
> addressed by changing "to specify the default language of any XML 
> character data sent over the stream" to "to specify the default 
> language of any human-readable XML character data sent over the 
> stream". Ted may want to further comment on this.

That makes sense. Does the following text address the concern and
capture the consensus from last October?

   4.2 Stream Attributes

   ...

   o  xml:lang -- An 'xml:lang' attribute (as defined in Section 
   2.12 of the XML specification [1]) SHOULD be included by the 
   initiating entity on the header for the initial stream to specify
   the default language of any human-readable XML character data it 
   sends over that stream. If included, the receiving entity SHOULD 
   remember that value as the default for both the initial stream 
   and the response stream; if not included, the receiving entity 
   SHOULD use a configurable default value for both streams, which 
   it MUST communicate in the header for the response stream. For 
   all stanzas sent over the initial stream, if the initiating entity
   does not include an 'xml:lang' attribute, the receiving entity 
   SHOULD apply the default value; if the initiating entity does 
   include an 'xml:lang' attribute, the receiving entity MUST NOT 
   modify or delete it (see also xml:lang (Section 8.1.5)).  The 
   value of the 'xml:lang' attribute MUST be an NMTOKEN and MUST 
   conform to the format defined in RFC 3066 [15].

   ...

   8.1.5 xml:lang

   A stanza SHOULD possess an 'xml:lang' attribute (as defined in 
   Section 2.12 of the XML specification [1]) if the stanza contains 
   XML character data that is intended to be presented to a human 
   user (as explained in RFC 2277 [14], "internationalization is for 
   humans"). The value of the 'xml:lang' attribute specifies the 
   default language of any such human-readable XML character data, 
   which MAY be overridden by the 'xml:lang' attribute of a specific 
   child element. If a stanza does not possess an 'xml:lang' attribute,
   an implementation MUST assume that the default language is that 
   specified for the stream as defined under Stream Attributes 
   (Section 4.2) above. The value of the 'xml:lang' attribute MUST be 
   an NMTOKEN and MUST conform to the format defined in RFC 3066 [15].