On Wed, 15 Jul 1998, Jul,Erik wrote:
>
> > -----Original Message-----
> > From: Simon Cox [SMTP:[log in to unmask]]
> >
> >
> [snip]
>
> This paper is posted at
> http://www.agcrc.csiro.au/projects/3018CO/metadata/dc_tf/type_1.html
>
> The text is as follows:
>
> [snip]
>
> > These can be defined and used as follows:
> >
> Just some general questions regarding the definitions:
>
> Is the DC community or the responsible WG making reference to any
> existing definitions? If so, could they (or, shouldn't they) be cited?
>
> If the DC community or the responsible WG is not refering to other
> existing definitions, why not?
>
> How will we know that we have a good definition? How will we know to
> what extent our definitions agree or disagree with usage in other
> resource description communities? Upon what authority will the DC
> definitions be based? And why do we imaging that the definitions will
> be accepted by other resource description communities?
>
> My concern is that definitions that are not demonstrably anchored in an
> awareness and appreciation of existing lexicographies may, for lack of
> such awareness *or* proof of such awareness (through citation), fail to
> achieve the level of acceptance needed to achieve the stated goal of
> "interoperability."
>
> Perhaps a subsequent draft of the document in question can orient itself
> and its users to the lexical traditions to which it adheres or from
> which it departs.
>
> --Erik
>
> Erik Jul
> [log in to unmask]
Erik:
Your message brings up an important point, and I've had to think about
this for awhile. However, I do think you could make the same argument
about the definitions of each element of the Dublin Core. Should they not
also refer to some established definitions? Obviously they've taken a life
of their own. The reason being, of course, that they are intended to be a
"lingua franca" for different communities and encompass more general
concepts. I would say the same for the Resource Type list in that it is
also intended to be a general list not tied to particular applications or
interests.
I looked at the Resource type list in terms of definitions in one well
established resource description community, MARC. By the way, this element
was very difficult to map to a single MARC field because there were so
many different ones concerned with type of resource at various levels of
granularity. The two obvious places to look were Leader/06 for type of
record (based on type of material), but that has too many distinctions
being drawn (values include: language material (i.e. text), printed music,
manuscript music, cartographic material, manuscript cartographic material,
projected medium, nonmusical sound recording, musical sound recordding,
2-dimensional non-projectable graphic, computer file, kit, mixed
materials (i.e. archival collections), 3-d artifacts, manuscript language
material). The more useful was 008/26 Type of computer file that has the
following: numeric data, computer program, representation (i.e. graphic),
document (i.e. text), bibliographic data, font, game, sounds, interactive
multimedia, online system or service. Again there are many more choices
than the shorter Resource Type list. That makes their definitions unusable
because the MARC definitions have to make distinctions between these
various other types. In addition Type of computer file is supposed to
cover only electronic resources, and I think (?) we now realize that DC
may be used for non-electronic.
It may be possible to adopt some of the wording in these definitions, but
not to wholeheartedly embrace them because they don't cover the same
concept. If someone else has an idea of what other resource description
schemes we should evaluate for definitions, we could consider it. I bring
up MARC mainly because it is intended to accommodate all types of material
and the other established schemes I know of are mostly subject specific.
However, it supports a level of complex search and retrieval (among
other things) that we wouldn't expect from the short list of 15 Dublin
Core elements.
Rebecca
|