Well, now I have to disagree.
First if you are agreeing with Aaron that "we should encourage...people
to use DC [the elements], even when it goes beyond the simple uses that
we intend it for", then this is very much in opposition to what I
believe. In my earlier message I raised the danger of the RDF value
method implicitely allowing the use of DC elements as a dumping ground
for any arbitrary sub-graph contaiing information related in any way to
the semantics of the element. Now you are agreeing that we might even
encourage such.
In your reply to my earlier message regarding this danger you stated:
stu >>> I agree that this is a significant danger. One way to avert
this
danger is to enforce the "appropriate literal" convention *AND* to
promote
only the use of formally defined schemas as appropriate subgraphs. One
pof
the roles of the DCMI Registry should be a repository of such schemas or
pointers to schemas maintained by others << stu
I hasten to point out that "promoting only the use of formally defined
schemas as appropriate subgraphs" does nothing to control anything. I
can right a formal schema for any nonsense I want.
Furthermore, I think we disagree about modular metadata. Placing
arbitrary sub-graphs within the value space of a dc property arc is in
my mind the antithesis of modular metadata. It is in fact promoting a
view that I strongly reject in my D-lib paper - that is, that the DC
elements might serve as the root of more complex resource descriptions.
As stated in the paper, this shouldn't be done since the simple resource
centric model and the sloppiness of the elements is incapable of
expressing such complex descriptions.
I feel like we're looping back with an RDF slight of hand to a place
that has led us into considerable trouble - the advocation of a model
where the DC elements can serve as containers for arbitrarily complex
information.
Sorry, but I'm being misinterpreted here.
Carl
> -----Original Message-----
> From: Weibel,Stu [mailto:[log in to unmask]]
> Sent: Wednesday, February 21, 2001 3:11 PM
> To: [log in to unmask]
> Subject: Re: vCard as a structured value strawman
>
>
> Aaron Swartz writes:
>
> I see DC as a foundation upon which much can be built. We should not
> discourage people to use DC, even when it goes beyond the
> simple uses that
> we intend it for. Instead, we should encourage such
> development, as it will
> strengthen the semantic meaning of documents.
>
> This view is very much in keeping with direction that I would
> like to see DC
> develop.
>
> RDF, as an architecture for modular metadata, makes this a
> realizable goal,
> and this goal need not conflict with the simpler instances of
> the standard
> that some want to adhere to. All we need is a convention for a simple
> parsing rule that makes it easy for applications to ignore
> what they don't
> understand. This is the essential benefit of the proposal that Carl
> advanced earlier today.
>
> stu
>
|