State of language-tagged literals.

In the last TC, David asked for a summary of the language tagged 
literals issue (ISSUE-71).



Richard's proposal is [1] which introduces classes rdf:LangString and 
rdf:Text.

We can then write rdfs:range statements about literals.

rdf:Text = xsd:string union rdf:LangString

xsd:string and rdf:LangString are disjoint.


The value of language-tagged literals
is a pair <Unicode string, language tag>.

The technical term “language-tagged string” is used for rdf:LangString 
literals.



In addition, I am suggesting giving a single datatype to language-tagged 
strings.

* All literals have a datatype.

* Some literals represent languages-specific strings;
   they have a string and a language tag.

* Literals that represent non-language-specific values.
   they do not have a language tag
   they have a lexical form.

One of Richards concerns was how this all works into documents, 
particularly around the names for concepts as, at the moment, 
language-tagged strings have to handled specially in text.

I'm suggesting there are:

+ literals
+ language-tagged string

just like

+ literals
+ numeric literals


rdf:PlainLiteral remains a datatype - it's usefulness is giving a 
lexical form to language-tagged strings, and also by forming a datatype 
that covers the class rdf:Text.  The fact that the value space overlaps 
with rdf:LangString and xsd:string is unimportant.

 Andy

[1] Richards proposal:
http://lists.w3.org/Archives/Public/public-rdf-wg/2011Jul/0048.html

[2] owl:Real --
http://www.w3.org/TR/owl2-syntax/#Real_Numbers.2C_Decimal_Numbers.2C_and_Integers

Received on Tuesday, 9 August 2011 15:27:22 UTC