>I have been given a problem from a greenstone user at the local University;
>They are using greenstone to manage a e-Titles collection; this is a
>pretty standard collection of PDF files (student access only so I
>can't show you)
>Links to the titles are included on the Library catalogue so students
>can click straight to the relevant item in the greenstone collection.
>The problem is they want to a periodical as a path to all the issues
>of a particular journal:
>- fixed link to a periodical (from library catalogue)
>- fixed link to individual issues (from library catalogue)
>- full dc metadata for issues
>- full dc metadata for each periodical.
>- periodical to show all associated issues.
>NOTE: 'fixed link' in the sense we don't want to have to keep updating
>the URL in the library catalogue.
>a (i) I suggested a special metadata field for the journal title, (or
>issn) - and just link to the classifier (via AZCompactList) -
>unfortunately the classifier changes on a rebuild (requiring a change
>to all the links in the OPAC)
>a (ii) the same as a(i) but link to a 'constructed' search like the
>one generated to bring up all the messages by a particular sender in
>the 'Greenstone Archives' collection
>b) as a 'web site' ; having a page(html) for each periodical with
>links to each issue(pdf) - a lot of work making pages and a web site.
>c) Multiple collections; create 'periodicals' collection and seperate
>collections for each periodical- then just add each collection as a
>web page and add metadata to taste, using a template collection for
>each periodical. (and making it 'public false' so the list of
>collections is not cluttered)
>I am thinking option (c) is best as it gives me a fixed URL for each
>issue and peridical - by using an assigned OID - The generated hash
>for the 'periodical' colelction (dynamically generated) homepages is
>unlikely to be the same.(but it should be fine for individual PDFs)
I've discussed this with Katherine and here's what we think we would do:
- Use "-OIDtype assigned" with import.pl. This uses dc.Identifier (you
could change this if necessary) metadata to form the HASH ID, so you can
be sure to get a fixed OID. In your case you would use something like
the ISBN number.
- Use your option a (ii) to get the periodical pages. This seems to be a
simple way of getting the information you want, with a fixed URL, and
without getting involved in classifiers or multiple collections. One
complication is that you need to make sure the issues are returned in
the correct order from the search -- you would use "boolean" searching
and add "-sortmeta <IssueNumberMetadata>" to import.pl to achieve this.
Hope this helps,
All the best,