Re: Multiple Semantic Web Languages (was: Re: reifying variables)

From: Pat Hayes (
Date: 02/24/04

  • Next message: Mike Dean: "Joint Committee telecon today 2 March"
    >Peter wrote:
    >>  > I argued long and loud in the W3C WebOnt working group about problems that
    >>>  using the RDF syntax caused.  This argument didn't go anywhere, so I gave
    >>>  in and created a partial solution for OWL.
    >Sandro asked:
    >>Do you remember why the WG disagreed with you?
    >Peter answered:
    >>Because all Semantic Web languages have to be same-syntax extensions of RDF.
    >Yes, I must support this. The *only* argument for many WebOnt 
    >members to accept/put up with the RDF syntax for OWL was political 
    >pressure (perceived or real) from W3C.
    Many, maybe: but not all.
    >Ian Horrocks wrote:
    >>Talking about cost (which you do below), probably the biggest cost in
    >>building OWL tools is dealing with the incredibly cumbersome RDF
    >I must also support this. This is not just a grudge-carrying claim 
    >from Ian, but substantiated by reports from almost all the 
    >contributors to the "OWL implementation experience" workshop Ian and 
    >I organised just before ISWC'03 in Florida. I asked all the 
    >contributors (all implementors of OWL tools) what they loved most 
    >about OWL and what they hated most about OWL. The RDF/XML syntax was 
    >by far the most frequent answer to the 2nd question. Subsequent 
    >discussion revealed estimates of up to 80% of development time for 
    >various OWL tools being devoted to the RDF/XML syntax. The 
    >underlying reason is the lack of a natural mapping from OWL to 
    Well, Im sure this is accurately reported, but the case needs to be 
    stated more carefully. First, is it really obvious that the triples 
    mapping is the primary problem, or is it the RDF/XML syntax which 
    gives everyone nightmares? But OK, there is certainly no natural 
    mapping from OWL-DL to triples: worse, the -DL restrictions cannot 
    even  be characterized in ways that can be located in the 
    OWL/RDF-triples syntax, so that the best that JENA2 can do, for 
    example, when given an illegal OWL/RDF graph which is not -DL legal, 
    is to reject it as unsuitable without being able to say why. 
    However, the blame for this should not be laid entirely at the door 
    of RDF: these issues would arise even if OWL were delivered in an XML 
    syntax directly modelled on the AS.  They arise from the non-local 
    nature of the artificial and illogical restrictions necessary to fit 
    within the DL subcase.  It is easy to use OWL constructs within RDF, 
    preserving their primary meaning, and with very few syntactic 
    complications. (The use of hd/tl lists to encode more-than-binary 
    arities in a relational format is admittedly ugly, but familiar: it 
    has been used in many pre-SW systems.)
    >W3C always says they value the opinion of "real" people who build 
    >"real" tools. This is it.
    >  ----
    IHMC	(850)434 8903 or (650)494 3973   home
    40 South Alcaniz St.	(850)202 4416   office
    Pensacola			(850)202 4440   fax
    FL 32501			(850)291 0667    cell

    This archive was generated by hypermail 2.1.4 : 02/24/04 EST