Quoting Brunella Longo <brunella.lo...@yahoo.com>:
The key question is: what does qualify a dataset as a dataset?
What is a dataset unique characteristique? Is data encoded in a
certain predictable structure (this is the DC definition).
The RDA definition is:
Dataset - Factual information presented in a structured form.
What
describe its nature or its structure (cartographic, financial,
patent, etc) does not matter and should perhaps be put in the
representation area of the description of the resource and not
together with the rdacontent designation name. In sum I would
recommend not to say that a cartographic dataset is a cartographic
dataset. Just say that it is a dataset.
That depends on what you want to do with your metadata. Metadata is
not a theoretical description of the world, it is a description that
needs to support particular applications (which is why we can't have
just one metadata scheme). Oftentimes, added sublevels exist because
of some particular need. It may be that there is a need to provide
particular properties for cartographic datasets that are not valid for
the class of datasets in general. Then again, the use of cartographic
dataset may merely have historical antecedents that make its
continuing use necessary to be compatible with legacy data. If you
look at RDA, there are a lot of specific cartographic content types.
The cartographic community clearly felt a need for these at some point.
In properly structured metadata, a cartographic dataset *is* a
dataset, and you can make use of it at either level of granularity. At
the same time, if you need to include particular information about a
cartographic dataset that is not valid for, say, a statistical
dataset, you do not want to assign those attributes at the wrong level.
With a structure like:
dataset
cartographic dataset
scale
azimuth
longitude
latitude
you can "dumb up" to dataset for your own purposes, while the
cartographic folks can enjoy the detail they wish to encode.
kc
But this is just a preliminary reflection and hypothetical solution,
not at all conclusive.
There is an increasing fascinating and systematic production and
distribution of datasets in many fields (I quoted as example
<http://openuplabs.tso.co.uk/datasets> but there is plenty of others
- factual, numeric, cartographic, financial, chemical etc - almost
always supposed to be downloaded via WWW services and then archived,
accessed and used through desktop software, smartphone applications
or server databases / repositories even if we cannot exclude other
generations of carriers).
Now, the (relative) peril I have seen is that cataloguing these
resources one could be tempted to apply some sort of "special"
solution that RDA and ISBD deserved to cartographic datasets also to
other type of datasets - for instance "statistical dataset". This
because uses and users seem demanding the distinction of a context
of use as extremely valuable: you don't access cartographic and
patents datasets with the same software! whereas it seems to me
that for cataloguing purposes in a Web environment the nature, scope
and other attributes of datasets and their uses should be treated
via other elements of the resource description - in sum it is not
necessary to say what type of dataset is a resource once you have
identified it as a dataset.
Brunella Longo
7 New College Court
London NW3 5EX
T +44 (0)20 72095014 (home) - +44 (0) 75 49921488 (mobile)
http://www.brunellalongo.info (http://www.brunellalongo.it)
--
Karen Coyle
kco...@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet