Ivan’s private site

January 13, 2007

BibTeX in RDF

Filed under: Semantic Web,Work Related — Ivan Herman @ 11:53

I just blogged elsewhere on BibSonomy, an alternative to del.icio.us that also includes bibliographical references. The core terminology used by the system (for bibliographical information) is based on BibTeX but entries can also be dumped to other formats. See, for example, a portion of my references in BibTeX or RDF.

However… which RDF terminology for BibTeX? Unfortunately, there are several around. BibSonomy uses the http://swrc.ontoware.org/ontology# namespace (usually referred to as “SWRC”), defined by ontoware. The bibtex2rdf tool by the Simile Project at MIT uses the http://simile.mit.edu/2006/11/bibtex# namespace (unfortunately, there is no namespace document). Incidentally, Steve Cayzer blogged yesterday on his usage of yet another tool, also called bibtex2rdf, that uses the http://www.edutella.org/bibtex# namespace (again, no namespace document). And, of course, there is also an example in microformats that could be used with either GRDDL or RDFa.

And all these vocabularies are different, unfortunately. For example, the ontoware version (used by BibSonomy) has a major flaw in my view: authors are just listed as individual property-values, without enclosing them into a Seq or (better) in a collection. Anybody who has ever co-authored a paper knows that the order of contributors is important. The usage of VCard, Dublin Core, FOAF, etc, is also different from one tool/vocabulary to the other. Pity… Maybe it is worth for the community to sort this out?



  1. BibTeX is only suitable for the hard sciences. We can do better, and I’ve started to address a more comprehensive model at .

    You’re right about the contributor issue; it is probably the most difficult one to model with the right balance of flexibility, rigor, and terseness. The solution I came to (not yet at that URI, but hopefully soon) in conversation with people like Dan Brickley and Ian Davis is to treat contributors as resources, with optional properties that indicate order. This seems the simplest and most consistent approach. It’s SPARQL-friendly, for example.

    Comment by Bruce D'Arcus — January 15, 2007 @ 19:33

  2. There are nearly 100 RDF ontologies that are based on or incorporate Bibtex. See


    It would be good to select/create one to use as a reference standard. Then we could migrate to it or at least define mappings to it.

    Comment by tim finin — January 15, 2007 @ 20:24

  3. In fact, it is also useful if some re-use of vocabularies is done. For example, in SwetoDblp we’ve used FOAF for declaring that someone is a person. The schema we have incorporates mappings to other vocabularies, using owl:equivalentClass or owl:equivalentProperty. I’ll be glad to see a vocabulary becoming ‘standard’ so that we could just use it right away.

    Comment by boanerges aleman meza — January 16, 2007 @ 3:18

  4. HubMed’s RDF export uses a mix of DC, PRISM, FOAF and BibTeX. It’s not perfect, but seems to work ok… Here’s an example: http://www.hubmed.org/export/rdf.cgi?uids=17216305

    Comment by alf — January 16, 2007 @ 11:28

  5. Ivan,

    Check out our openacademia project at [1], our web-enabled, semantics-based publication management system. We are using a BibTeX serialization that is RSS 1.0 (so also RDF), uses the ontoware vocabulary, but fixes the major flaw you mention (loss of the order of author names, and also editor names). The format is called BuRST and described at [2].


    [1] http://www.openacademia.org
    [2] http://www.cs.vu.nl/~pmika/research/burst/BuRST.html

    Comment by Peter Mika — January 16, 2007 @ 13:33

  6. […] April 16, 2007 at 5:07 am · Filed under bibliography management, RDF, Semantic web BibTeX in RDF « Ivan’s blog […]

    Pingback by BibTeX in RDF « Aman’s Blog — April 16, 2007 @ 7:21

  7. Beginning with ISWC2006, we also use SWRC (with some additions as suggested by Peter Mika) for the metadata corpora of ISWC and ESWC. The plan is to have all conference metadata available through one service, which is currently being set up at http://data.semanticweb.org


    Comment by Knud Möller — May 29, 2007 @ 13:19

  8. […] openacademia.org project to generate RSS feeds of the documents in their RDF triple store. See also Ivan Herman’s blog post on this […]

    Pingback by Rinke Hoekstra » Converting BibTeX to RSS feeds: bib2rss.pl — January 15, 2008 @ 9:38

RSS feed for comments on this post.

Blog at WordPress.com.

%d bloggers like this: