Delasalle, Jenny wrote:
>
> I do have three potential solutions to my metadata creation problem:
>
> 1) Employ more cataloguing staff
> 2) Minimise the processing and cataloguing by requiring academics to do
> it
> 3) Employ a technical wizard to automate stuff.
>
> All of these could work in conjunction with a fourth option, which is to
> compromise the thoroughness of the metadata record, which could be a
> temporary approach to be addressed by any of the other methods at a
> later date. Or indeed it might turn out to be a permanent solution. The
> point of detailed metadata records is what functionality they support,
> either in terms of search/reporting within our own repository, or
> interoperability with others. This is hard to judge because we're trying
> to look into the future at what technology might enable us to do with
> our metadata.
Interesting thread. I can't give any wise words, I can let you know what
we do.
To answer your original question. We haven't really timed it but I would
guess between 5-10 minutes per item.
Less if no full text. much more if something like a book chapter where
the copyright will require emailing academics/publishers.
We check copyright (mainly relying on Sherpa Romeo), and will make small
improvements to the metadata, adding DOI/ISBN, fixing little problems
(why do people type with caps lock on).
When we set ours up, it was somewhat implicit that it should require
practically no resources after a year of operation, so of course
decisions were made based on this.
Though, like I think many Universities, the RAE and REF pilot have
helped convince people they need a good database of publication output,
with good metadata. I'm hoping the repository can be the solution for
this, and appropriate support provided. (at which point our metadata
standards will probably move a little from the 'self-deposit sloppy' end)
I've quoted Jenny's email at the top, as I think it sums up my thoughts
far better that I could have said myself (especially regarding the
reasons for metadata: management info, and system interoperability). We
use LOC subject headings, because it seemed like one way that a third
party system could harvest records for a particular subject from many
Universities. But the confusion and effort it causes, and the lack of
any tools/websites making use of it, means I'm on the verge of removing
them.
So, taking the ideas you mention above, and from what others have said:
- getting academics to put at least some information in has to be a good
thing. a techy can help make it simple and perform basic checks
- I confess to not knowing what exactly SWAP requires, so can't picture
how much information you are entering.
- What metadata is good for internal management information/REF? What
information is very likely to be useful for system interoperability
(both other IRs and other campus systems)? Perhaps anything which falls
in to neither can be abandoned? The second is harder to answer. We all
run the risk of trying to predict the future, and I wonder sometimes if
we should keep it simple until we know what's useful, by which time
automated tools will probably have been developed to do it for us!
- At a more general level, I found the "make live quick, add extra
metalib later" approach an interesting idea. I can imagine a specific
'metadata editor' functionality which makes this efficient: a screen
which shows all fields for a record on one view, and once done, takes
them straight to the next record which needs editing.
- food for thought: librarything have a large database with records with
minimal need for editors. Can we learn from them?
I was going to make this a quick email, apologies for the length.
Regards
Chris
|