Hi Bernard,

For me at least, the DCAM is probably best understood as the data model component of a DCAP -- a formal model that should be independent of its implementation but ultimately still machine-actionable. As you point out, Both The DCAP and the DCAM exist in an incomplete state as formal DC recommendations or standards and this, probably more than anything else, has limited their uptake. It's hard to evangelize and put into practice a relatively complex set of recommendations that are merely almost done. Plus the utility of the DCAM, and in particular it's distinction from RDF, has been questioned by its creators over the last few years and this has helped move people in other more concrete and stable, but perhaps less effective modeling directions.

Some of these issues are what we're trying to address by taking up the discussions again.

I would like to think that a more robust and evangelized DCAP/DCAM standard would be of great use outside of the DC community. Perhaps people would find uses for the DCAM outside of the context of a DCAP, and it might be interesting to try to collect use cases if we can figure out a way to explain it sufficiently clearly to do that. But it doesn't seem to me to be a good place to start.

Tom's very helpful posts from earlier today should also help explain where the discussion is currently headed.

Btw, LOVBot is a wonderful acronym!

Best,
Jon

On Thursday, February 2, 2012, Bernard Vatant <[log in to unmask]> wrote:
> Hi all
>
> I've been trying to follow the discussion on DCAM revision, and I must confess I've been totally confused so far on what the conversation was all about until this post from Pete, which is the first I fully understand and agree with.
> To try and understand more other viewpoints, I have a slightly provocative question : what is the use of the current DCAM so far, outside the DCMI standards?
> If I look from the RDF vocabularies ecosystem, the answer is : nobody AFAIK :(
> Compare the reuse of DC Terms and DCAM at
> http://labs.mondeca.com/dataset/lov/details/vocabulary_dc.html
> http://labs.mondeca.com/dataset/lov/details/vocabulary_dcam.html
>
> Of course the current content at http://dublincore.org/2010/10/11/dcam.rdf is minimal, but one would think that the very generic class dcam:VocabularyEncodingScheme would be re-used here and there. One would expect for example things like
>
> skos:ConceptScheme  rdfs:subClassOf  dcam:VocabularyEncodingScheme
> See discussion http://lists.w3.org/Archives/Public/public-esw-thes/2010Jan/0007.html
>
> Is my question coming from a narrow-minded view from RDF land? If yes change my question to :
> Who *else" outside RDF land and DCMI standards cares or should care about DCAM? And why?
>
> Maybe a preliminary answer to this question would help me understand where this debate is bound to.
>
> Thanks for your time
>
> Bernard
>
>
> 2012/2/1 Pete Johnston <[log in to unmask]>
>
> Hi all,
>
> I'm a bit (OK, very!) confused about this analogy between DCAM and SKOS.
>
> To me, SKOS has two components:
>
> - a model of (a part of) the "world" as made up of concepts, concept schemes, lexical terms etc, which have certain attributes and certain relationships between them
> - an RDF vocabulary (or two if you distinguish base SKOS and SKOS-XL) for use in creating RDF graphs/triples to describe that "world"
>
> SKOS is quite generalised so it can condition how we choose to model our "worlds" in other domains (e.g. do I model my "places" as SKOS Concepts with broader/narrower relations or as spatial things with contains/is-contained-by relations? And so on)
>
> But using SKOS doesn't determine/change the nature of my data structures, or the "lens" I apply to those data structures; it only changes my "world" structures: using SKOS I'm still squarely within the framework of RDF graph and triple data structures. SKOS Concept Schemes and Concepts are just more "things" in the "world", but in terms of how my data about those things is "packaged", SKOS Concept Schemes and Concepts are treated exactly the same as any other thing (a foaf:Person, a bibo:Document, a dcmitype:Collection etc etc etc).
>
> But - with its notions of Description Set, Description and Statement - DCAM does introduce new data structures, or at least (as I prefer to try to think of it) a new "lens on", a new way of looking at and referring to parts of, the RDF graph/triple structure.
>
> In contrast to SKOS, with DCAM, it's not a question of looking at "the world" in a different way. Whether I think of my data as an RDF graph or a DCAM Description Set (or as both, depending on how I'm looking at it!), my "world" is still the same: it has foaf:Persons who author bibo:Documents that are about skos:Concepts that are in skos:ConceptSchemes.
>
> Rather with DCAM, I'm looking at the structure of my _data_ in a different way.
>
> So I'm afraid I'm struggling to grasp the significance of comparing the DCAM to SKOS - at least at the level that comparison seems to be being applied in these discussions. I understood Andy's mention of SKOS on 05/01
>
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;dc1738b9.1201
>
> to be about the practical usefulness of SKOS, the fact that it addresses a requirement that people have ("how do I represent my thesaurus using RDF?"), not saying that DCAM was something "similar in nature" to SKOS.
>
> Further on in that thread, Kai said on 09/01:
>
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;1fc1d387.1201
>
>> RDF is not only defined for the representation
>> of metadata, it is so abstract that at the same time, it allows for instance the
>> definition of ConceptSchemes in SKOS. And if there is a need for the
>> definition of a ConceptScheme, I argue that there is a need for the definition
>> of a DescriptionSet, too.
>
> I think this is where I got lost :)
>
> (To me), a SKOS Concept Scheme is just another thing in my "world" (alongside a FOAF person etc), another thing to be named with a URI and described in my data, my graph, using RDF triples.
>
> But a DCAM Description Set is "a thing in my data", not in my "world". Sure, I could name and describe it (just as I could name and describe an RDF graph or an RDF triple) but my main "use" of the Description Set notion is as a way of structuring my data.
>
> So, from my perspective, I can't help feeling that an SKOS Concept Scheme and a DCAM Description Set are very different things, and I'm struggling to grasp why comparing them is useful.
>
> I'm not saying it isn't useful, just that, right now, I don't "get it" :)
>
> Pete
>
> Pete Johnston
> Technical Researcher
> Eduserv
> E: [log in to unmask]
> T: +44 (0)1225 474323
> F: +44 (0)1225 474301
>
> --
> Bernard Vatant
> Vocabularies & Data Engineering
> Tel :  + 33 (0)9 71 48 84 59
> Skype : bernard.vatant
> Linked Open Vocabularies
>
> --------------------------------------------------------
> Mondeca                             
> 3 cité Nollez 75018 Paris, France
> www.mondeca.com
> Follow us on Twitter : @mondecanews
>

--
Jon

I check email just a couple of times daily; to reach me sooner, click here: http://awayfind.com/jonphipps