[standards-jig] UPDATED: JEP-0071 (XHTML-IM)

Iain Shigeoka iain at jivesoftware.com
Wed Sep 17 05:21:33 UTC 2003

On Tuesday, Sep 16, 2003, at 12:43 US/Pacific, Peter Saint-Andre wrote:

> Version 0.8 of JEP-0071 (XHTML-IM) is now available. The changelog is 
> as


We recently stumbled across an XHTML issue regarding entities. I think 
it will be common for people to assume since they see 'XHTML' in the 
standard, that the XHTML predefined entities will be valid and even 

However, use of XHTML entities in an XMPP stream is not valid XML. It 
breaks a MUST rule of the XML spec (section 4 of xml 1.0) regarding 
entities that aren't defined before use. in addition, it further 
violates a 'MUST NOT' XMPP specification by implicitly bringing in new 
character entities (section 9.1 of XMPP core).

I was wondering if an implementation note should be added warning 
against the use of XHTML predefined entities (beyond the 5 XML entities 
that are defined for XML (< > & " '). Instead 
implementors should be encouraged to encode characters that aren't of 
the predefined 5 using a character reference instead (see section 4.1 
of the XML 1.0 specification). These encodings are of the form &#; 
where the # is the hex value of the character from the ISO/IEC 10646 
character set and is valid XML in XMPP, and valid XHTML.


More information about the Standards mailing list