And the follow-up, for completeness.
---------- Forwarded message ----------
From: Riina Vuorikari <riina.vuorikari eun.org>
Date: 23-mar-2007 11:05
Subject: Re: [DC-EDUCATION] Vocabulary criteria: discussion
wanted!
To: DC-EDUCATIONjiscmail.ac.uk
All,
I would like to send a contribution on behalf of Silvia
Panzavolta
(s.panzavoltaindire.it) on the Thesaurus discussion.
She was
involved, through INDIRE (It), in building the LRE (ex-ETB)
Thesaurus. I think her point on the use of international
standards is
rather important. Thanks for your interest and this
discussion on our
Thesaurus work. Don't hesitate to contact us for any further
questions.
Riina
Find it below:
The standard for the construction of a multilingual
thesaurus that
was used is ISO 5964 and ISO 2788 by ISO (see http://
www.collectionscanada.ca/iso/tc46sc9/standard/5964e.htm),
which is an
international organization. NISO standards, on the contrary,
are only
approved by Americans and are national standards, though
they can be
adopted in other countries. They for sure have a lower
degree of
general consenus. Apart from that, the standard referred to
in the
discussion, Z 29.19, is not published online and does not
seem to
have been approved (at least this is the info published on
the site).
Moreover, among the NISO standards, none is about the
construction of
multilingual thesauri.
Back to the LRE thesaurus and ISO standards:
The relations are compliant with the standards mentioned
above. The
associative relation (USE/UF relations) is, of course, the
"hardest
to define" as ISO 2788 expresses. I cite: "It
refers to the link
between two terms [!] which do not form an equivalence set,
nor can
they be organised as a hierarchy, yet they are associated in
common
usage to such an extent that user who refers to one of those
terms in
an index or thesaurus should be directed to the other
term". In the
example mentioned, since we did not introduce the terms
"foster
child" nor "adoptive parents" we thought that
users may think of
those terms as entry points too, and thus we decided to
create such a
relation (USE/UF). Of course, other USE/UF relations could
have been
determined, but at that time we thought that the 2
non-descriptors
(NDs) were the minimum set to be provided.
The comment on "mixing classes of person with
concepts" when using
the associative relation is totally natural and compliant
with the
standards since they are such because of a "strong
association in
common use". I would not say that in common usage, a
user is not
allowed to associate library and librarian, for instance,
still being
them very different kinds of "objects" and
"concepts".
As for the systematic display, it is well defined according
to the
standards, since the ISO 5964 admits several features as for
the
systematic display of a multilingual thesaurus. I cite:
"thesauri
differ very widely in their approaches to the relative roles
and
functions of these two sections [alphabetic and systematic]
and in
the arrangement and kind of relational information provided
in each."
Finally, the distinction between concepts and terms is not
exact,
according to me, since you only can express concepts through
terms
(so it is impossible to be so strict in their distinction,
they are
two different level of analysis: semantics and linguistics).
Of
course, when selecting a certain amount of terms to be
included in a
controlled and limited vocabulary as a thesaurus is, the
constructors
of the thesaurus use a concept-oriented methodology
(semantics). For
instance, "home" is both a concept and a term,
with a variety of
synonyms that can be regarded as equivalents of a same
concepts (i.e.
house, apartment, flat). Again, according to the scope of a
thesaurus, this might be even be questionable. A thesaurus
about
architecture, for example, would probably adopt different
criteria.
Hope I clarify a little bit the logics behind our thesaurus
work.
All the best,
Silvia
Dr. Silvia Panzavolta, Documentalista Senior
s.panzavoltaindire.it
Dipartimentimento 'Documentazione, Didattica e Formazione' -
Sezione
'Documentazione'
I.N.D.I.R.E. (ex BDP)
Istituto Nazionale di Documentazione per l'Innovazione e la
Ricerca
Educativa
Palazzo Gerini - Via Buonarroti, 10 - 50122 Firenze
Tel: +39-0552380427 - Fax: +39-0552380514
http://www.indire.it
On 21 Mar 2007, at 15:50, Fredrik Paulsson wrote:
> Diane,
> The LRE thesaurus is actually not my project - even
though I was
> involved several years ago. I can also see the problems
pointed out
> by Mikael and you. The main reason for pointing out
this work is
> that there is a lot of work put into it by a lot of
people, and
> that it might be used as a basis for further
development that might
> take to a point where it will met the DC requirements.
To my
> knowledge there are no other multilingual thesaurus
around for
> education and the LRE could maybe be used as a
"starting point". It
> could at least be used as a point of contact for
consensus - as it
> has been used before.
>
> However, I think that Riina Vuorikari at EUN is a more
suitable
> person to answer for the LRE and its future development
than I am 
> Regards
> Fredrik
>
>
> Diane I. Hillmann wrote:
>> Fredrik:
>>
>> I agree with Mikael that this is an impressive
piece of work, and
>> I appreciate the pointer to it. Besides the concern
noted by
>> Mikael that it seems to be more
"word-based" than "concept-based"
>> I would also note that it doesn't seem to follow
standard
>> thesaural guidelines for relationships (for
instance, in Z29.19
>> 2005). For instance, for the concept
"adoption" you assert the
>> following relationships:
>>
>> *UF */ adoptive parents
>> /*UF */ foster child/
>> *RT* parents
>> These seem problematic to me, in that they seem
somewhat random,
>> mixing concepts like adoption with classes of
person without any
>> indication of what kind of relationship is
asserted. I looked also
>> at your systematic display, and note that these
relationships are
>> not reciprocal in your thesaurus, most likely
because it is not
>> concept based.
>>
>> Now, I may very well be missing something, and I
hope you'll
>> enlighten us if we've made some incorrect or
inappropriate
>> conclusions. I also note that the distinction
between word lists
>> and concepts may not be considered as important in
the IEEE LOM
>> world as it is in Dublin Core, and there are
certainly important
>> efforts in progress to bridge that gap.
>>
>> Thanks again for pointing us to this work, and I
hope you'll keep
>> us updated as it evolves.
>>
>> Regards,
>> Diane
>>
>>> fre 2007-03-16 klockan 14:04 +0100 skrev
Fredrik Paulsson:
>>> > One educational, multi-lingual thesaurus
that might be of use
>>> is the
>>> > ETB. Developed within the European
schoolnet. It contains
>>> educational
>>> > terms in 14 different languages.
>>> > http://insight.eun.org/ww/en/pub/insight/interoperabil
ity/
>>> learning_resource_exchange/metadata.htm
>>> > /Fredrik
>>>
>>> While this is an important and thorough piece
of work, I believe it
>>> unfortunately fails to meet the criteria set
forth by Diane &
>>> Sarah -
>>> thanks to the fact that it's a word-based
thesaurus, relying on
>>> language-specific definitions of each word.
>>>
>>> Thus, I think it's unfortunately not very
useful in the DC-Ed
>>> application profile
>>>
>>> Please correct me if you think I might be
wrong, either in
>>> understanding
>>> ETB (or the current incarnation) or in applying
the draft criteria!
>>>
>>> /Mikael
>>>
>>> >
>>> > Diane I. Hillmann wrote:
>>> > > Marie-Claude:
>>> > >
>>> > > I certainly think we should include
vocabularies in a variety of
>>> > > languages (and bilingual
vocabularies)--we'll certainly do so if
>>> > > they're suggested to us. However, as
you know, it's
>>> difficult to find
>>> > > them without knowing already where
they are, or having them
>>> pointed
>>> > > out as you've done.
>>> > >
>>> > > I took a stroll through the
vocabularies in your registry and
>>> there
>>> > > didn't seem to be any specifically
addressing education. Are
>>> there
>>> > > any in progress, do you think? I
know Canada has had a
>>> number of
>>> > > educational initiatives over the
years, and it seems odd to
>>> me that
>>> > > there were no vocabularies developed
for those.
>>> > >
>>> > > Diane
>>> > >
>>> > >> Diane,
>>> > >>
>>> > >> Anything about the languages of
the vocabularies listed for
>>> use in the
>>> > >> Education AP?
>>> > >>
>>> > >> In the Government of Canada (GC),
we went through a similar
>>> process to
>>> > >> determine which vocabularies to
use within our governmental
>>> context
>>> > >> (so no
>>> > >> particular focus on the education
domain), and we came up
>>> with the
>>> > >> following list of criteria:
>>> > >> http://www.collectionscanada.ca/information-managemen
t/004/
>>> index-e.html
>>> > >> (under "How to
register").
>>> > >> In summary: must be
"controlled" / publicly available /
>>> bilingual /
>>> > >> maintained by a trusted
authority.
>>> > >>
>>> > >> To be really useful to the
largest audience possible, the DC-
>>> Education
>>> > >> registry of vocabularies should
include vocabularies in
>>> other languages
>>> > >> than English. Of course, a local
community could always
>>> translate an
>>> > >> appropriate vocabulary into its
own language. However, this
>>> is not an
>>> > >> ideal
>>> > >> solution. In the GC, when we
develop controlled
>>> vocabularies, our best
>>> > >> practice is to develop both
English and French versions in
>>> parallel,
>>> > >> instead of developing in one
language and translating into
>>> the other.
>>> > >> This
>>> > >> is a best practice, in reality it
is not always followed. As
>>> French and
>>> > >> English languages function
differently, it is advisable not
>>> to simply
>>> > >> translate. It has happened to us
that a concept in one
>>> language was non
>>> > >> existent in the other. Just a
thought.
>>> > >>
>>> > >> MC
>>> > >>
>>> > >> Marie-Claude Nancy Côté
>>> > >> Agente principale des normes /
Senior Standards Officer
>>> > >> Patrimoine canadien / Canadian
Heritage
>>> > >> Culture canadienne en ligne /
Canadian Culture Online
>>> > >> 25, rue Eddy, 3ième étage / 25
Eddy Street, 3rd floor
>>> > >> Gatineau (Québec), K1A 0M5 /
Gatineau, QC, K1A 0M5
>>> > >> Tél : / Tel: (819) 934-1295
>>> > >> Télécopieur : / Fax: (819)
934-1293
>>> > >> marie-claude_n_cotepch.gc.ca
>>> > >> www.pch.gc.ca/progs/pcce-ccop
>>> > >>
>>> > >>
>>> > >>
>>> >
>>> >>
>>> > >> "Diane
>>> I.
>
>>> >>
>>> Hillmann"
>
>>> >> > >>
>>> <dih1CORNELL.EDU
To
>>> > >> >
DC-
>>> EDUCATIONJISCMAIL.AC.UK > >>
Sent by:
>>> > >> DCMI
cc
>>> > >>
>>> Education
>
>>> >>
>>> Community
> >> Subject
>>> > >> <DC-EDUCATIONJIS
Vocabulary criteria:
>>> discussion
>>> > >> CMAIL.AC.UK>
>>> wanted! >
>>> >>
>>> >
>>> >>
>>> > >> 03/11/2007
>>> 06:09
> >>
>>> PM
>
>>> >>
>>> > >> Please respond
>>> to >
>>
>>> "Diane I.
>
>>> >>
>>> Hillmann"
>
>>> >>
>>> <dih1CORNELL.EDU
>
>>> >>
>>> >
>
>>> >>
>>> >
>>> >>
>>> > >>
>>> > >>
>>> > >>
>>> > >> Folks:
>>> > >>
>>> > >> Sarah and I have been discussing
regularly the tasks and
>>> progress of the
>>> > >> DC-Ed Application Profile. We
hope to make regular postings
>>> outlining
>>> > >> issues we see as we get into this
work and continue thinking
>>> about
>>> > >> what the
>>> > >> Application Profile should do for
us.
>>> > >>
>>> > >> One of the main tasks has been
collecting appropriate
>>> vocabularies,
>>> > >> primarily for Type (focusing on
learning resources) and
>>> Instructional
>>> > >> Method, though other
educationally oriented vocabularies
>>> will also be
>>> > >> welcomed. See the wiki at:
>>> > >> http
://dublincore.org/educationwiki/Vocabularies for a
fuller
>>> > >> discussion of
>>> > >> our task and what we've collected
so far.
>>> > >>
>>> > >> An issue that has come up
involves exactly what we're
>>> expecting from
>>> > >> vocabularies. As we know,
vocabularies can run the gamut
>>> from simple
>>> > >> lists
>>> > >> of terms (sometimes used for pull
down lists) to fuller
>>> vocabularies
>>> > >> with
>>> > >> lead-in terms, definitions, and
appropriate encodings. When
>>> it comes to
>>> > >> listing vocabularies within the
AP, what are we looking for?
>>> > >>
>>> > >> Herewith is an attempt to propose
some criteria for
>>> inclusion that
>>> > >> focus on
>>> > >> reusability. This is not to say
that we will not include
>>> > >> vocabularies that
>>> > >> do not comply with all these
factors, but we would certainly
>>> prefer
>>> > >> those
>>> > >> that do:
>>> > >>
>>> > >> 1. The vocabulary is openly
available. While there is a
>>> place for
>>> > >> licensing
>>> > >> in the vocabulary world, most
educational organizations
>>> aren't able to
>>> > >> choose vocabularies that require
licensing, either because
>>> of financial
>>> > >> constraints or institutional
size/resources. Also
>>> problematic are those
>>> > >> vocabularies only available via
specialized tools, or
>>> 'hidden' behind
>>> > >> complex websites.
>>> > >>
>>> > >> 2. The vocabulary includes
definitions. Particularly because
>>> the DC-Ed
>>> > >> community is international,
assumptions that words are used
>>> the same
>>> > >> everywhere are problematic. This
is true even when the words
>>> are
>>> > >> English,
>>> > >> and relatively common: the
context of their use by their
>>> primary
>>> > >> community
>>> > >> may affect relationships and
boundaries not obvious to others.
>>> > >> Definitions
>>> > >> assist mightily with this
cross-cultural understanding, and are
>>> > >> essential
>>> > >> for effective mapping between
vocabularies.
>>> > >>
>>> > >> 3. A standard encoding is
available for the vocabulary. This
>>> criteria
>>> > >> implies that the vocabulary is
available as a whole, not
>>> just term by
>>> > >> term
>>> > >> from a search interface. This is
important for those
>>> attempting to
>>> > >> develop
>>> > >> new tools, needing to be able to
import a vocabulary into
>>> that tool,
>>> > >> rather
>>> > >> than using copy/paste or
referring to an ever changing web
>>> page. The
>>> > >> availability of a standard
encoded version also imples a
>>> much more
>>> > >> machine-oriented view of how
vocabularies will be used and
>>> re-used.
>>> > >>
>>> > >> 4. URIs have been assigned to the
terms and/or the terms are
>>> registered.
>>> > >> This is a bit of a holy grail,
and few vocabularies are in this
>>> > >> category.
>>> > >> Including this criteria orients
us properly in the right
>>> direction and
>>> > >> hopefully moves us closer to that
ideal. If you're curious
>>> to see
>>> > >> what a
>>> > >> registered vocabulary with URIs
looks like, see:
>>> > >> http://metadataregistry.org/vocabulary/show/id/24.html
or
>>> > >> http://metadataregistry.org/vocabulary/show/id/11.html
>>> > >>
>>> > >> In both cases, you can click on
'concepts' to see a list of the
>>> > >> individual
>>> > >> concepts in the vocabularies,
then click on the concepts
>>> themselves
>>> > >> to see
>>> > >> the details and relationships.
Encodings can be viewed by
>>> using the
>>> > >> buttons on the bottom for RDF or
(at the top level
>>> vocabulary only)
>>> > >> the XML
>>> > >> Schema.
>>> > >>
>>> > >> So, please do think about these,
suggest others, protest,
>>> etc. If we
>>> > >> can
>>> > >> come up with something that's
fairly agreeable across the
>>> board, we
>>> > >> can add
>>> > >> it to the wiki.
>>> > >>
>>> > >> Regards,
>>> > >> Diane
>>> > >> --
>>> > >>
>>> > >>
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
*
>>> > >> Diane I. Hillmann
>>> > >> Research Librarian
>>> > >> Cornell University Library
>>> > >> Email: dih1cornell.edu
>>> > >> Voice: (607) 387-9207
>>> > >>
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
*
>>> >
>>> >
>>> --
>>> <mikaelnilsson.name>
>>>
>>> Plus ça change, plus c'est la même chose
>>
--
Juan Rafael Fernández
http://people.of
set.org/jrfernandez/
|