Re: Object Memory Modeling XG and RDFa

Dear Manu Sporny,

appologies for the late reply - we had to take care with the OMM XGR in 
the last week. And thanks a lot for the careful review of our proposal - 
we've discussed it and plan to come up with a detailled reply 
(...including a revised proposal) in the next days!

Best regards,
   Alexander Kröner

Am 10.10.2011 04:28, schrieb Manu Sporny:
> On 10/05/2011 05:32 AM, Alexander Kröner wrote:
> > I've wondered if you found the time to have a closer look on our
> > examples, or if you require additional information about the OMM. We
> > would like to add some information about this OMM aspect to the
> > OMM XGR, so your feedback would be of special value for us.
>
> Apologies, but until this point, I had not been able to look at it. My 
> response is below, hope it helps in some small way.
>
> On 09/26/2011 10:42 AM, Alexander Kröner wrote:
>> the W3C Object Memory Modeling Incubator group (OMM XG, [1])
>> experimented with an RDFa-based encoding of the Object Memory Model
>> (OMM, the main result of this XG). An example, which was validated with
>> W3C RDFa Distiller and Parser, is located here:
>>
>> http://www.dfki.de/~haupert/files/omm_rdfa.html
>
> A couple of things on a quick look through the source code:
>
> 1.  You should be using RDFa 1.1. Instead of xmlns:, use prefix, like
>     so:
>     prefix="omm: http://www.w3.org/2005/Incubator/omm/elements/1.0/"
> 2.  Your declaration of xsd is not correct, it should be:
>     prefix="xsd: http://www.w3.org/2001/XMLSchema#"
> 3.  Your OMM vocabulary isn't human or machine readable, you should
>     public a vocabulary document at:
>     http://www.w3.org/2005/Incubator/omm/elements/1.0/
> 4.  I suggest that you do not version your vocabulary with minor
>     release information unless you intend to publish minor releases,
>     and if so, those minor releases shouldn't break backwards
>     compatibility in which case the minor release number doesn't
>     matter. In other words... just use major numbers:
>     http://www.w3.org/2005/Incubator/omm/elements/1/
> 5.  You don't need to express the version number of your vocabulary,
>     your vocabulary URL above does that. Don't do this:
> <span property="omm:version">1</span>
> 6.  If your primary ID is a URL, it should be expressed as one, so
>     instead of this:
> <span property="omm:primaryID">http://example.com/p1</span>
>     do this:
> <a rel="omm:primaryID" href="http://example.com/p1">...</a>
>     You should be doing this with all of your URLs
> 7.  This whole Header, Table of Contents, Block stuff in OMM just
>     does not seem like the right way to model this information. You
>     have a graph data structure available to you... don't treat it
>     like tape-based storage data structure. :)
> 8.  omm:uri - this term is suspect... why do you need it? IRIs are
>     first-class citizens in RDF. Do this instead:
> <a rel="omm:additionalBlock" href="http://URL_OF_BLOCK">...</a>
> 9.  omm:type is suspect as well. Instead of this:
> <span property="omm:type">omm_http</span>
>     do this:
> <span typeof="omm:Http">...</span>
>     or something to that effect.
> 10. This is redundant:
> <span property="omm:date"
>           datatype="xsd:dateTime">2011-01-31T08:12:50+02:00</span>
>     of encoding: <span property="omm:encoding">ISO8601</span>
>     xsd:dateTime is an ISO8601 formatted date, you don't need to
>     specify that it is again in the next triple.
> 11. omm:hasTag looks a bit like a mess - it looks very complicated and
>     open ended.
>
> Other than that, the RDFa markup looks fairly good. I think the 
> biggest issue is the OMM vocabulary itself - it seems to be 
> over-modeling. If you don't make the vocabulary simpler, the chances 
> of uptake are reduced greatly.
>
> Instead, of trying to think of it in terms of Headers and Tables of 
> Contents and Blocks... just think of modeling the data as you would a 
> real object - just add Event objects to the main object and timestamp 
> them. That said, I don't really know what use cases you need to 
> support - but the vocabulary seems quite complex for the intended use 
> (and it's not documented, so I can't understand how each vocabulary 
> term is meant to be used).
>
>> We first discussed this experiment with our colleague and RDFa WG member
>> Sebastian Germesin, and now wanted to share our experiences. And of
>> course we are interested in feedback concerning our attempt, e.g., if
>> the OMM structure could be expressed in a more accurate or simpler way.
>
> The RDFa isn't the problem, IMHO... it's the vocabulary. I've made a 
> few suggestions above that I think could help, so maybe you could 
> respond to those and we could see how we may go from there.
>
>> Once we agree on the representation, we might also proceed with a
>> discussion of our “user experience” with RDFa, if you are interested in
>> that.
>
> Yes, we are very interested in hearing about the issues that you may 
> have hit when using RDFa. I think that due to the complexity of the 
> OMM data model, the RDFa is going to look very complex indeed. To give 
> you an example of what a simpler vocabulary could do, you may want to 
> look at the schema.org examples on this page:
>
> http://linter.structured-data.org/examples/
>
> -- manu
>


-- 

Dr. Alexander Kröner --------- Intelligent User Interfaces Lab
DFKI GmbH  Campus D3 2  Stuhlsatzenhausweg 3 66123 Saarbrücken

Phone  +49.681.85775.5395
Fax    +49 681.85775.5021

http://www.dfki.de/~kroener/
--------------------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern

Geschaeftsfuehrung:
Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender)
Dr. Walter Olthoff

Vorsitzender des Aufsichtsrats:
Prof. Dr. h.c. Hans A. Aukes

Amtsgericht Kaiserslautern, HRB 2313
--------------------------------------------------------------

Received on Wednesday, 19 October 2011 08:34:45 UTC