One week has elapsed since the position paper on DC.Type
from the Type & Format working group was posted.
I have detected the following issues raised:
------------------------
1. Off-line and Surrogates
The 1:1 principle is still not fully understood,
in particular in regard to _off-line_ resources.
We need to develop more examples illustrating the use of
DC metadata for both physical objects _and_ digital surrogates of these.
------------------------
2. Data
There is a concern that "data" is not sufficiently distinctive.
The definition could be improved by adding the adjective
_structured_ in order to differentiate it from
_unstructured_ things that are "text".
Perhaps change the token to "dataset".
------------------------
3. Image and other symbolic notations
Questions about where _maps_ and _musical_notation_ fit
exposed the need to improve the definition and examples for "image".
Perhaps use "symbolic visual representation".
This type should also not be restricted to 2D.
Perhaps change the token to "graphic".
------------------------
4. Video
Is video OK as "image/graphic" ...?
What is video?
i. format rather than type information ...?
ii. "moving pictures" - in which case type "image/graphic" is fine
iii. "moving pictures with sound" - this is the compound/mixed issue:
use an additional DC.Type="sound" if the sound is a major aspect.
(NB. lots of video (particularly on the web) is silent.)
Lots of the classifications are not strictly orthogonal anyway.
The parsimony principle argues against adding types willy-nilly.
---------------------
I've posted a new version of the DC.Type position paper
revised by me to incorporate some of these changes at
http://www.agcrc.csiro.au/projects/3018CO/metadata/dc_tf/type_2.html
---------------------
5. Previous work
Meanwhile, Erik Jul has observed that the list of Types is not
properly grounded in an analysis of previous related work
in this area. This has two implications:
(i) the taxonomy may not be fully robust
(ii) there is a risk of a credibility gap wrt some of our target communities
Rebecca and I have fessed up and thought about this a little.
Here's my attempt at a summary of the points raised:
- "Type" is not alone in DC in proposing semantics without
explicitly indicating antecedents
- the DC community has frequently appeared to reinvent such a wheel,
but this is not necessarily totally wierd since a cross
community "lingua franca" needs to encompass more general
concepts than those of any specific community
- this DC simple list is unusual in that it does attempt to cover
such a wide range with such a small number of terms;
besides, it has stood up to most empirical tests over a number of months now
- some specific comparisons have been attempted (MARC)
but was found to be rather difficult since there was a
correspondence with multiple MARC fields,
frequently dealing with different levels of granularity
- the problem will almost certainly be less stark for more
refined DC with external vocabularies used explicitly through "schemes"
Meanwhile:
Erik:
> there must be a handful of reference
> sources, dictionaries, standards whose entries we we examine, compile,
> and compare. MARC is one. There are others. First step would be to
> compile a list of potential resources. Sounds like a good reference
> question for a librarian somewhere.
Rebecca:
> With that said, it is probably worth the effort. The
> problem of course is, who has time? Not a good excuse, I know. Maybe we
> need a game plan and each of us take various groups to see what they've
> done.
I think (hope) Rebecca and Erik are working on it ...
--
__________________________________________________
Dr Simon Cox - Australian Geodynamics Cooperative Research Centre
CSIRO Exploration & Mining, PO Box 437, Nedlands, WA 6009 Australia
T: +61 8 9389 8421 F: +61 8 9389 1906 [log in to unmask]
http://www.ned.dem.csiro.au/SimonCox/
|