On 4 May 2008, at 21:06, Jewel Ward wrote:
> Margaret Alexander wrote me off list and told me I should cite my
> study, so here it is:
> Ward, J. (2004). Unqualified Dublin Core Usage in OAI-PMH Data
> Providers. OCLC Systems and Services, 20(1), 40-47. Available via: http://www.ibiblio.org/jewel/foar/research/mp/jward-ocls_sands-apr2004.pdf
> .
If I understand it correctly, your conclusion is that we not using
enough of DC - and from your data it looks like the Language,
Publisher, Format, Rights, Contributor, Source, Coverage and Relation
terms that are missing. It is interesting that the parts of DC that
are used heavily are precisely those parts that are useful for
resource discovery (title, description, creators, date etc). Since
most resource discovery is in fact done via Google, we are back to the
original discussion point!
You indicate in your paper that further study should be directed at
the reasons for the missing DC terms - did you do any work on this?
> N. Dushay and D. Hillmann, "Analyzing Metadata for Effective Use and
> Re-use," DC-2003: 2003 Dublin Core Conference, Seattle, September
> 2003.
> http://dc2003.ischool.washington.edu/Archive-03/03dushay.pdf
This paper concentrates on a method of visualising metadata encoding,
rather than the metadata contents themselves.
> do i think the OAI-PMH framework or was dead?
I wonder whether the question should be "Is OAI-PMH dead, or still
undergoing a very protracted labour?". OAI doesn't deliver metadata
magically (unless the metadata already exists, in which case it
certainly can't be blamed for any deficiencies!) and repositories are
still getting to grips with issues about metadata - what metadata to
collect, how to use it for in-house reports, how to display it on
repository pages etc. What OAI has done is to put metadata on the
agenda for a wider range of users, but "How To Share It WIth Others"
has become a non-issue when Google does all our heavy lifting.
> the lack of shareable, quality metadata has been a barrier to using
> the framework to "its fullest extent", as the lack of quality
> metadata has limited the kinds/types/amounts of services that can be
> built on it.
Or is it the other way around? It is difficult to create and maintain
a global information service! There's a lot of infrastructure to get
in place (e.g. reliable network crawling and local storage) not to
mention trying to find a sustainability mechanism to keep it running
for more than 2 years. Sure, dodgy metadata doesn't make it any
easier, but who is going to invest in metadata improvements without
payback for doing so?
Perhaps if we were to build an academic social networking service that
capitalised on OAI feeds, then we might get a sudden rush to improve
the situation!
> yes, there have been hurdles to overcome, but it has been 5-6 years
> since the first formal studies came out on the lack of shareable
> quality metadata, and that has given some time to improve on what is
> out there...and that hasn't happened. Rachael Heery mentioned to me
> at the JCDL in 2003 that she had done an an informal study at some
> point in the 1990s (unpublished) in which she found similar results
> to my study published in 2003. in other words, this has been a
> known problem for a while, even if the results have been more
> informal and anecdotal.
There are a number of repository sites in the UK that have over the
last 24 months invested a huge amount of effort to clean up their
metadata for the Research Assessement Exercise. It would be VERY
informative indeed to look at how an external driver like the RAE has
impacted their metadata by comparing it to IRs which have not been
used by the RAE. And also to look at how whether that metadata is
optimally exposed in OAI.
> does that mean i think metadata is dead? no. just that there is a
> known need for improvement, but given the limited resources of
> libraries, that is not likely to happen in an optimal manner.
We have of course been talking about several different things: correct
use of OAI, correct and full use of DC and quality metadata expressed
in DC and transported by OAI. Quality metadata may be expressed in DC
and transported by RDF, and I look forward to greater integration
between the W3C's view on the Linked Data world, and the Open
Archiving perspective.
>> Organisations do not (on the whole) implement PMH, any more than
>> they implement SMTP. They install an EPrints/DSpace/Fedora server
>> like they do an Exchange server. Their users might provide
>> unsatisfactory metadata, but then they might send unsatisfactory
>> email too :-)
> whether or not one "implements" the OAI-PMH depends on what software
> you use; not everyone uses ePrints/DSpace/Fedora.
> in which case, they "implement" the protocol b/c it is not built-in,
You're right about my repository-facism. Please excuse me! But we
ought to check the figures - I was under the impression that most
people used a predefined OAI package for the development environment
of their choice.
---
Les
|
