[Zope-CMF] Re: Metadata Toolz

Erik Lange erik@mmmanager.org
Thu, 10 Oct 2002 23:59:22 +0200


At 01:55 PM 10/10/02, Tres Seaver wrote:
>On Thu, 2002-10-10 at 05:54, Erik Lange wrote:

<snip>

>For that we need to extend the
> > metadata namespace we use, whit a namespace that handles time-indications.
> >
> > This extension can be defined with dublin core, so dublin core should -
> > IMHO - always be the base of all metadata.
>
>No, I don't think so.  DC should be the "lingua franca" schema onto
>which other schemas can be mapped, with loss of fidelity.  It is *not*
>the Right Thing (TM) for all cases.

Hmm... I can't see where we disagree ?

When I say "base of all metadata", I mean that you in some cases may not 
use any of the elements in dublin core to describe your resource, but 
you'll use dublin core to describe other metadata standards who's elements 
you're using instead of the DC-elements... so DC will at least always 
function as "lingua franca" between the various metadata standards, even if 
no DC-elements are used. Right ?

> > Dublin core's "has.part" and "is.part.of" elements, will make it possible
> > to describe wich clips are related to which physical datafiles and vice 
> versa.
>
>Those aren't elements;  they are qualifiers on the Relation element:

Okay, you're right of course - forgive me my bad semantic - "the relation 
element and it's qualifiers, will make it possible to describe wich clips 
are related to which physical datafiles and vice versa", was what I meant...

Actually, a better example is the relations between content elements in 
composited/compound presentations/documents... the relation element and 
it's qualifiers is essential for "true" compound content handling... IMHO.


>   http://www.dublincore.org/documents/dces/#relation
>
>   http://www.dublincore.org/documents/dcmes-qualifiers/#relation
>
>DC also has a (redundant) Source element (Source is nothing more than a
>special, qualified Relation, where the qualifier would be 'Derives
>From', or some such).

Yes - except that in our case where Source would be the "masterfile", in 
which a given media clip is located, the element doesn't contain redundant 
metadata ;-)

 From DC:

Source element - Definition: A Reference to a resource from which the 
present resource is derived. Comment: The present resource may be derived 
from the Source resource in whole or in part.



>Right now, content in the CMF doesn't know about the Relation element,
>because unlike the other elements it is *complicated*;  it essentially
>*requires* qualifiers (like Date) in order to make sense.

Yes - it's *very* tricky stuff... that's why I tried to simplify my 
explenation and mixed elements and quallifiers... but you're right - it's 
important to use the correct terms, so we all now what we're talking about 
- thanks for the corrections.

>   A sane
>implementation is also likely to require something like Ken Manheimer's
>"organization objects" proposal, at least if relations are based on
>anything other than pure containment.

Hmm.. not sure I've read this - got a link ?

Regards,
Erik