RE: Example of coordination with DDC

Andrew, 
>I'm not sure I agree, maybe I misunderstood the above, but MARC21
>classification does provide the 765 field (SYNTHESIZED NUMBER 
>COMPONENTS) to describe the pieces of a synthesized number.
<http://www.loc.gov/marc/classification/eccdbild.html#mrcc765>

thx...this is my favourite subject!  :-) although hardly a SKOS issue
I am well aware of 765 and this field is textual and poorely structured
information about how the number is derived which is how it happens in DDC 
(base number and extension) - and does not allow for differentiation 
between different ways numbers are combined in analytico-synthetic 
classifications. It helps in searching but does not allow access to meaningful 
parts of notational strings and global changes in the classification heading 
field (150). This means that 765 is useless when it comes to data management 
and maintenance and impedes the creation of faceted interface. 

>Having just looked at the data in the Dewey editorial system, it
>appears that the editors have been adding the components of a 
>synthesized class since at least Edition 20, circa 1996.

Exactly. Well explained in:
MITCHELL, J. S.  (1997) "Challenges facing classification systems 
: a Dewey case study", Knowledge organization for information retrieval 
: proceedings of the Sixth International Study Conference on 
Classification Research, London, 16-18 June 1997. The Hague : 
FID, 1997. (FID 716), 85-89.

So it looks like to me that the approach from MARC 765 field 
does not help much in accessing classification data ;-)

>Being the person who has been looking at modeling the DDC in SKOS, I'm
>not sure that SKOS directly needs to define classes and properties for 
>expressing the components of a synthesized class.  SKOS tries to use the 

I don't know to which extent a proper vocabulary exchange format 
such as SKOS is really relevant for DDC users and if at all... 

I can only argue for the importance of such a format for e.g.
UDC which is normally distributed to users as a database
file and which they use to build or support their IR systems,or in
creation of their own vocabulary (thesaurus or subject heading system).
Also, there are enriched, multilingual UDC data, mapped to different vocabularies
that has been developed for years in library and other IR systems and that 
may be put to a wider use through formats like SKOS. 

It seems silly to me that a standard has to be built in such a
way that it is suitable for system that does not really need it and is 
less suitable for system that can make use of it?

>> DDC is simple enumerative classification with largely non-expressive  
>> notation which is used as text string to 'mark and park' books. 
>I think the Dewey editors would disagree with this statement...

This is not a criticism of DDC. It is about
functionality of a system and requirements based on it. 
Classifications are built around purpose
Auxiliary tables i.e. synthesis started to be introduced more in Dewey 
only in 1970s. Compared to analytico synthetic classification 
DDC is still 'simple' and it is largely enumerative - which is 
why it is more popular in the environments where classification is 
used in a simple way. This quality is exactly 
the reason why DDC is more suitable and more widely used in larger number of libraries 
(mostly public) - this was also the argument why it was more suitable for e.g. Renardus
gateway (according to T. Koch)
I doubt that anyone or editors of DDC can object much to the fact that DDC is
a) created and used primarily for library shelf arrangement
b) not known to be popular in information retrieval systems (it is simply
not specific enough for indexing of scientific or research papers)
c) sadly not used even in the library OPACs to the extend it
could or should be (I can cite here around 50 references to
support this and some of these written by Dewey editors but I suggest you look
at the research reported by Cochrane or Karen Markey Drabenstott)

Aida

Received on Thursday, 3 August 2006 11:55:30 UTC