Print

Print


Andy,

Am 06.01.2012 18:11, schrieb Andy Powell:
> So, (looking at
> http://wiki.dublincore.org/index.php/DCAM_Revision_Tech) you
> currently have 3 reasons for why people should invest their time in
> understanding and using the DCAM:

I have to admit that I merely wrote that section as a kind of 
placeholder, but also to foster a discussion and finally get an idea 
about our motivation, which I consider very important. So these goals 
for sure aren't very elaborated, but I am happy that the discussion 
started now :-)

> --- cut ---
>
> 1) Provide a way to publish metadata on the Semantic Web. 2) Provide
> an abstract model for metadata applications. 3) Help to bridge the
> (terminological) gap between the Semantic Web and the metadata
> community (Dublin Core Community)
>
> --- cut ---
>
> 1) doesn't seem overly compelling to me since I assume that most
> people will (rightly) consider that RDF allows them to do that
> (without need for the DCAM).

Indeed, but RDF is very general (abstract) and questions remain how to 
describe metadata in RDF and how to represent the concepts we are 
talking about in DCAM. The metadata itself is plain RDF, of course. But 
I assume that people would like to frame it in same way, e.g. by 
defining DescriptionSets.

> 2) may be of value in limited cases (though I question that this is
> what the DCAM effort is trying to do - assuming that "metadata
> applications" is intended to be interpreted in its broadest sense).
>

For me this is a very practical goal. It allows developers to create 
data models that are able to handle all kinds of data that adhere to the 
DCAM.

> 3) is part of the original rationale for the DCAM (and thus is of
> questionable widespread value, as I outlined in my previous message)
> though I would tend to change the last part to "... between the
> Semantic Web and the Dublin Core community (and hence other metadata
> communities)".

I thought that DCMI as a whole is such a bridge so DCAM would be one 
building block of this bridge.

> If 1) is important, then you need to spell out (for me at least) what
> DCAM brings that RDF does not already have.

Notion of a Description, Notion of a DescriptionSet, well defined ways 
to talk about both and the contained (RDF) metadata.

> If 2) is intended as written, then I question the scope of this
> activity, since "metadata applications" is such a broad space, and
> DCMI's ability to deliver it.
>

We have already the DCAM, the question still is what to do with it. I am 
inclined to embed it completely in RDF, both to clean it up and to 
remove any ambiguities that exist between DCAM and RDF. But that's only 
one approach and clearly depends on the goals. I think it's doable, 
BECAUSE of the embedding: we do not have to reinvent anything that's 
already available in RDF. I wonder, if it is then still useful as an 
abstract model for non RDF applications and how big the change is for 
existing applications that use DCAM, that's something we have to find out.

> 3) is OK (though, again... as written, the scope is far too broad to
> be realistic) but on its own is not enough to justify this effort.
>
> Apologies for seeming negative... my concern is that the DCAM served
> a purpose at a particular point in time (i.e. the point where DCMI
> was allowing multiple encoding syntaxes to flourish without a good
> understanding of what they each represented) and was primarily
> intended to help clarify DCMI's own thinking. One can argue about how
> well the DCAM served that purpose but the real point now is that we
> are no longer at that point in time. If DCMI wants to influence the
> wider sphere thru this work, then you (collectively) must articulate
> why anyone should sit up and take any notice. I would strongly
> suggest doing so before investing much intellectual effort in doing
> the actual work on a revised DCAM.
>
> Just my 2p... you are, of course, free to ignore me ;-)


I would never ignore you ;-) You are absolutely right and you raised 
valid points. I agree that the purpose of DCAM will (and has to) change. 
I think that something like DCAM is missing in RDF, maybe just because 
noone thinks that it is needed. 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.

Cheers,

Kai

-- 
Kai Eckert
Universitätsbibliothek Mannheim
Stellv. Leiter Abteilung Digitale Bibliotheksdienste
Schloss Schneckhof West / 68131 Mannheim
Tel. 0621/181-2946 Fax 0621/181-2918