[Fwd: Re: [greenstone-users] mixing Greenstone with DSpace, citesser, berry picking, thesaurus, name authority records, collocation]

From Chi-Yu Huang
DateMon, 06 Dec 2004 10:18:57 +1300
Subject [Fwd: Re: [greenstone-users] mixing Greenstone with DSpace, citesser, berry picking, thesaurus, name authority records, collocation]
Hi Karen,

Thanks for this.

Karen E. Medina wrote:
Dear Greenstone,

Greenstone is fantastic. And I want to thank you for adapting greenstone
to OAI and METS. And Greenstone3 looks fantastic. Greenstone has some
really fantastic parts to it -- which when combined with some other tools
would be really powerful. I am interested in combining some features from
Greenstone and DSpace, also creating internal links from documents to the
refered documents (kind of like how citeseer works), name authority
control, a thesaurus, and a few other tools to work in and between
collections. I know Greenstone was not intended for these purposes, but
part of the academic world is clamoring for institutional repositories.

A few questions: 
I'm kind of looking for advice on where to begin with any and all of

Access -- How hard would it be to index an item in Greenstone, yet have
DSpace control access to the full document? (DSpace has a strong emphasis
on permission to access).
You might want to include the URL address which contains the same verion of document in DSpace.  So you search using GS but when you want to retrieve the document GS is instructed to use the URL that obtains the same document from DSpace. In order to do that, the simplest way is to save the URL information in the metadata.xml file or use GLI (Graphical Librarian Interface) to manually add the URL information as metadata. Then, you need to change the format statement in the "collect.cfg" file.  Of course, the advantage of doing this through metadata.xml is that it should be possible to automate this.

Search support -- I'm still very interested in the berry-picking work from
early Greenstone. And supporting searchers even more.
We are currently looking at berry-picking stuff and hopefully will put it back to the next release version of Greenstone. 
Moving beyond Dublin Core -- Dublin Core is fantastic for homogenizing
collections, and necessary for OAI, but for local use, Greenstone is
flexible enough to create MARC-like records or even one's own metadata.
All these are important. DSpace's provenence leaves a lot to be desired. 

Adding Annotation -- allowing users to add annotations to the documents in
a Greenstone collection. This will probably be the simplest.
Greenstone uses GDBM (Gnu Database manager) to store metadata, classifier information, and document structure information. You can add new pieces of metadata to it. However, if would be deleted whenever you rebuild the collections. So if you want to define your own special annotations, the possible soultions would be:
    1.  Add annotations in the metadata.xml and put it into the import directory.
    2. Save annotation details as a separate database file and save that in the collection folder (say) not in the index folder.

Doing this would require either the extension of the existing action in the GS run time code, or a new action. A new action seems the better option at this stage.

Mixing with citeseer -- adding recommender type bits via citation.
Could you explain this idea a bit more clearer?

-Karen Medina

Hope these are useful!

Software developer
greenstone-users mailing list