Use of datatype rdf:HTML in multilingual void-Documents -- good or bad idea?

I turn to the list for some expertise and advice. 

While writing a void-Document for a linked data service I wanted to describe a dataset using some canned text containing an unordered list of features ("bullet points"). Since it is difficult to transport bullet points in an RDF literal I looked at other void descriptions and realized that some providers (it could have been DBPedia) uses html-code in combination with the datatype rdf:HTML in their dataset description. This works fine for the bullets, but makes it tricky to supply information in different languages since I cannot use a language code in combination with rdf:HTML, instead I need to work with lang="..." in the HTML content [2].

This boils down to the following question:
Is it a recommended good/best practice to describe datasets with html in order to be able to use some fancy features (such as lists) even if that makes it more difficult to find a description in the proper language (I guess it would be difficult to as a SPARQL endpoint to evaluate an html lang="..." attribute) or is it better to stick to plain text descriptions and use the standard ways of language identification, e. g. "..."@en-US or "..."@fr?

[1] http://www.w3.org/TR/rdf11-concepts/#section-html


Thanks in advance for any insight on this,

Lars

Received on Thursday, 2 May 2013 13:46:18 UTC