XML literals considered textual

[Some personal thoughts triggered by my understanding of Martin's messages]

Premise:  the primary role of XML literals in RDF is to allow 
text-with-markup to be captured as well as plain text.

Possible implications:

1. XML literals should be treated as similarly as possible to plain literals.

2. In particular, treatment of xml:lang applied to XML literals should be
the same as xml:lang applied to plain literals.

3. It may be desirable (but, I think, not essential) to distinguish between 
plain literals and XML literals in the abstract syntax.  (One possible 
alternative might be to treat all literals as XML, applying escaping as 
required to '&' and '<' in plain literals.  Personally, I think this is 
problematic.)

4. Possible use of XML structures to represent other datatyped values is a 
secondary consideration.   One might consider other datatypes using XML 
markup to encode their values, e.g. a complex number datatype with lexical 
forms like:
    <real>123.45</real><imag>54.321</image>
But, parseType="Literal" parsing cannot be used with datatyped literals, 
and it may be better to use RDF for making composite data structures rather 
than have two.

5. The intended purpose of XML literals should be made clear in the RDF 
specifications, if only for the benefit of future working groups.

I deliberately do not try to formulate any detailed conclusions at this time.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E

Received on Wednesday, 9 July 2003 04:53:13 UTC