RDF 1.1 string literal updates

In relation to the resolution of the RDF 1.1 WG on promoting simple
literals to literals of type xsd:string:

Personally, I quite like this change, as it has the potential of
making the systems I've worked with much easier. However, it opens up
problems for SPARQL 1.1.

As discussed, the RDF WG has not yet come to a decision about Plain
Literals with language tags, though a decision will be forthcoming.
The mere fact that the language tag is still a simple literal
indicates (to me, anyway) that a decision *must* happen. Given the
timing of these changes we now have to choose from several options:

1. Do nothing.
2. Incorporate the current RDF WG resolution on simple literals, and
nothing else.
3. Wait for a further resolution on Plain Literals before proceeding
to incorporate into SPARQL.
4. Loosen up the language around functions, etc, that operate with
strings, trying to leave things open for changes.

Are there other options?

Some comments on these options...

While #1 is tempting, it runs the risk of leaving SPARQL 1.1 subtly
incompatible with real-world requirements once RDF 1.1 is finalized.
On the other hand, it also guarantees backward compatibility, while I
suspect that an update to SPARQL may have difficulty with that.

#2 would involve some work, but the discussions so far make it seem
reasonable. However, it would seem strange to see *some* of the new
RDF 1.1 changes incorporated into SPARQL, but not others, especially
when these changes are all about string literals.

While #3 might provide the ideal outcome in terms of the spec, I don't
see this as workable given our time constraints. Not only may it take
a long time before a resolution is forthcoming, it will also take time
to work into the documents once we get it.

#4 is a compromise, but I really don't know how feasible it is.

Paul Gearon
Revelytix, Inc

Received on Tuesday, 21 June 2011 15:26:20 UTC