Rachel,
sorry for confronting your absolutely valid argument only
to make a few points of my own. Here are summary of two
practical issues

1) relation of local practices and national objectives

If the goal is to have national gateway of educational material this gateway
will have to use general knowledge system. So if you have a special collection
and use special classification such as NLM then you have to map your local
system to the general system, and this should be done through some kind
of mapping tools (authority file) which would not affect your everyday work
or indeed your local search gateway.

There is no general knowledge system that  will do for you the same as your
special system. They are fundamentally different in the way they treat disciplines
and arrange topics. You have your established practice and there is no need
to change that in order to expose your collection to the search on national
level.
In that respect on does not really care what is that general system and
what were the reasons for its choice.

2) About choice of general classification system for educational domain:

When you say 'Dewey is not suitable' - my answer is 'but of course it isn't
, it is not meant to be'.  This is notoriously obvious. If one goes and
chooses 19th century enumerative classification system created to arrange
shelves in public libraries - to do classification of potentially highly
granular,
multifaceted, heterogeneous and cross-disciplinary grounded learning resources
- this has only one meaning: the expectations and requirments for resource
discovery put on this system are extremly low. And this is how I understand
it: low expectations to make compromise which is likely to be accepted by
majority.

This means that ones does not really need classification system to relate
medicine to biology or biology to nursing, or to create relationship between
any two or more related disciplines in the way they may occur in the learning
material.
As this is exactly what Dewey is not meant to do.  And the problem is not
in the number of concepts in Dewey (20000), as they will keep growing, but
in the way these terms are structured and in the way this structure is locked
by the system.

It is, thereofore,  reasonable to expect that Dewey will be accepted by
majority in educational domain and, in my opinion, exactly because it has
to be paid for. Reasoning behind this is: if you have to pay for it and
if so many libraries are using it
must be good enough. Money one pays, can also sometimes be a guarantee in
making choice in the area in which there is not enough expertise to make
an objective judgement. Also, if classification system is free like BSO,
or Bliss  and not many libraries are using it - it must be something wrong
with it.
Does it matter that the truth is the opposite?
When these faceted systems were created (from sheer frustration with enumerative
classifications) - leading libraries have already had classification systems
in place, and one does not reclassify library collections just because there
are better systems. Also every new small library that opens has to follow
the example of older and bigger libraries. And especially in case with Dewey
which is supported by OCLC and which comes
for free with bibligraphic databases one buys and downloads into the library
system. I have heard of libraries with perfectly intelligent system that
moved to Dewey simply because this one comes with OCLC products and for
libraries that meant less
hassle. OCLC, in return, is adding terms to classification existing structure
to modernize it and is building tools on it which is finances by the users'
money.
(this is not the case with non-English speaking countries where OCLC bibliographic
database cannot be bought to describe local collection)

It would be ideal if educational community would take one of free, faceted
vocabulary and encode it using something similar to XFML  (eXchangeable
Faceted Metadata Language
www.xfml.org)  expose it on the Internet so people can create tools which
would extract index  terms, thesaurus etc.  and enable flexible and interactive
search and browsing interface, mapping to other vocabularies (even Dewey)
and building classification and automatic classification tools.
Such  a system would be much more in support of semantic web idea.

But things like this never happen, do they... So it looks like that valid
thinking here is to preserve local systems and find the way to map to Dewey
for the sake of interoperability.
Mapping has to be external and done only once and should not ask for continuus
effort.

Aida
-----Original Message-----
From: The CETIS Metadata Special Interest Group [mailto:[log in to unmask]]On
Behalf Of Rachel Ellaway
Sent: 24 February 2003 16:40
To: [log in to unmask]
Subject: Re: UKCMF questions



Aida (and everyone)


the problem is a semantic and ontological one. If the authors and consumers
of metadata are involved at the abstract end of use then a general appreciation
of the issues you put are likely. However if the users are practitioners
within well established communities of practice then (as I have observed
on a number of occasions) the arbitrary division that places their world
view in a, to them, inappropriate context  does a lot of harm to the confidence
in and sense of ownership of the metadata.


Re your PS: 600 is 'technology' rather than 'applied science' which would
have been more appropriate but its separation from biology remains a major
conceptual flaw.


It is also notable that Dewey is a commercial system that requires licensing
 from the OCLC Online Computer Library Center Inc. What have been the licensing
agreements with CETIS and the UK HE/FE communities regarding implementing
this centrally and locally?


I represent chalk-face users that may all too easily fall away from us if
our systems reflect antithetical world views to their own. However, in the
interests of progress and consensus we will let this one go for now, unless
anyone else wants to pitch in?


best


Rachel









>Rachel:
>may I only pick only on the proposal to use Dewey
>
>> 9.1 Purpose:  X4L value space for discipline: mandatory:
>> Dewey - this
>> was raised in Manchester at the JORUM meeting and the point that a
>> common system for all users is essential as a core 'glue' was well
>> argued by Lorna. However the Dewey is for some subject
>> areas (with my
>> medicine hat on) particularly dysphoric with the discipline's
>> worldview. For instance medicine is under 'technology' and not
>> 'science' and is in different categories from bioscience and other
>> health-related areas - this lack of conceptual sense is likely to
>> disenchant and disenfranchise those who have to negotiate it. Either
>> this is relatively hidden from users or a better system is going to
>> be needed. I don't have an answer but it is an issue that is
>> concerning us ... magic bullets anyone?
>
>This about a compromise.
>Any classification with small classification base such as Dewey
>has to 'subsume' large amount of disciplines, the whole
>of knowledge, under the very small number of categories
>in what is called a 'roof' or base of  classification.
>I understand that this may be a bit odd from your point of
>view dealing with medical science only and using probably
>NLM classification which works with alphabetical, non-decimal
>notation .... where on can start with 26 classes instead of 10.
>
>Every general classification has the problem of scientific
>and educational consensus when ordering disciplines.
>I would like to mention an example of broad knowledge structure
>that solved the problem of obsolete classification
>base by following, what is known as, integrative level
>in sequencing of sciences. This is only to illustrate
>that you were right complaining about Dewey's compressed
>structure.
>The system is called the Broad System of Ordering
>created by FID&UNESO in the seventies, to serve as a switching
>language between classification systems such as Dewey, LCC, UDC
>BC, CC etc.. However, although it is
>freely available it came too late to be widely applied in information
>services(you can have look at http://www.ucl.ac.uk/fatks/bso/), and
>too early to be properly supported by technology.
>I put last year the whole classification of around 10.000 concepts on
>the web (it only in text format for the time being).
>
>Hence, using Dewey or any other classification to produce map
>to the universe of knowledge has to be some kind of compromise in
>order to provide browsing features spanning the whole of
>knowledge.
>
>Why Dewey, and not some other more intelligent system?
>The answer is exactly the same as the one we get on the
>question 'But why Microsoft?'. Dewey has 100 years long
>history of being distributed as a part of a 'bibliographic packages'.
>So one has to forget how dumbed-down it is, simply because
>it has become an 'interoperable' solution.
>
>There are not much time or resources that implementors
>in educational domain can afford to waste pondering on
>classification systems. This is why there is a tendency to favour the least
>
>advanced systems such as Library of Congress Classification or Dewey.
>Kind of quick fix. Compared to LCC, Dewey can, indeed, strike as rather
>
>sophisticated knowledge organization tool (mind you, everything
>compared to LCC does).
>
>Education field does need some kind of general knowledge
>classification structure. If one has to object Dewey, that has to be
>based on its failure to satisfy some basic requirement in
>indexing educational material, rather than its failure to appear
>logical on the first level of division  - but this is an
>entirely different issue.
>
>Aida Slavic
>SLAIS,
>University College London
>
>P.S. your argument about medicine being under applied
>sciences would not be a good one, in this case. I hate to defend
>Dewey, but medicine is by all means an applied science and can't be
>classified under pure science such as chemistry, biology etc.
>Medicine is field that quite clearly applies chemistry, biology,
>physics etc. So it has to go together in 600 with technology.