On Fri, 2003-03-14 at 16:59, Dan Brickley wrote: > In an XML encoding of RDF, you could use XML entities to keep document > size down through text re-use. In an RDF implementation, database etc. you > are free to invent your own optimisations to avoid duplication of > textual content. I'm not convinced we _need_ extra machinery in RDF itself > based on this use case (though I can imagine it being handy...). Storing it is not the problem, but rather referring to it from within RDF. It cannot, for example, be reliably reified: if the textual content is modified, the reification is invalid. We have this problem when trying to store layout information for presentations of RDF graphs - there is no way to say "that literal", e.g, "that literal should be placed there" and at the same time allowing that literal to change. I do think the use of literals as subjects is a stronger argument, though ;-) /Mikael -- Plus ça change, plus c'est la même chose