On 4/16/12 8:14 AM, [log in to unmask] wrote:
> Karen
>
> The framework requires that a single value from each of the vocabularies
> forming a base category (for content or carrier) is used to underpin the
> high-level category. In your examples, "image" can have Sensory mode =
> "touch" as well as "sight", so cannot map to a single Sensory mode value;
> and "music" can have Sensory mode = "sight" (notated music) and "touch"
> (braille notated music) as well as "hearing". This is a powerful,
> fundamental feature of the framework, as it forces high-level categories to
> make the distinctions explicit
But at the same time, it makes it impossible to link data between
disparate systems. We're back to the question of rigidity of
definitions, much like we ran into with FRBR. It seems that it would be
more useful to define ISBD "image" as >= RDAONIXimage+touch or >=
RDAONIXimage+sight than to say that they are incompatible. I think that
we'll have more cases of making these kinds of reconciliations than we
will have of finding data from different sources that has exactly the
same definitions.
I guess I had assumed that the goal of the mapping was to ... well, to
map two different but similar sets of terms so that they could interact
in the big data soup. Maybe that's the next step? Because this seems to
be a comparison without compromises.
RDA has separate categories/terms for
> "still image", "tactile image", "three-dimensional moving image",
> "two-dimensional moving image", "performed music", "notated music" and
> "tactile notated music".
>
> The use-cases behind the framework included the GMD and SMD vocabularies
> used in AACR2, and the increasing difficulties in using them (and adding
> new terms consistently) as new forms of digital content and carrier exposed
> the semantic ambiguity in the "traditional" approach. This has been
> extensively discussed on the RDA-L listserv.
And I assume that the "solution" will be to create mappings between the
previous sets of terms (AACR2 mainly) and RDA. In fact, on the RDA list
right now this is being discussed because I wanted to see a comparison,
and people have (seemingly successfully) been able to do that because I
have received some responses that do that mapping.
I want to second what Barbara said:
>>> Wouldn’t it be a useful step at this "early" stage to harmonize ISBD
> and
>>> RDA as the JSC and publishers re-open discussions with the publishing
>>> community? Going in such different directions does not seem helpful to
>>> international standardization efforts, and it is not clear why ISBD
>>> chose to take such a divergent approach. Perhaps such harmonization
>>> discussion would enable us to jointly agree on a better approach.
I think "harmonize" is the key statement here.
>
> I think your examples of "image" and "music" can be considered use-cases
> for abandoning the GMD/SMD approach.
I'm not so much concerned that we need to abandon the GMD (although we
will with RDA) as that we now have a data format with conceptually the
same data in multiple fields using different values, with no defined
relationships between them. Creating those relationships between ISBD
and AACR with RDA will be needed if we are to transform our data to...
well, whatever it gets transformed to. Declaring that ISBD area 0 and
RDA/ONIX are different (and we know they are) doesn't solve my use case,
which is to have a lot of data from different sources play together well.
kc
--
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
|