Service Providers

Service providers who specialize in a topic provide useful functionality limiting the service to a specialized topic. For example, the Avano marine and aquatic sciences service provider (http://www.ifremer.fr/avano) clearly conveys what are the topic strengths of the service (marine and aquatic sciences.) Avano has a very useful Browse Archives feature, which displays the collections Avano draws from, links to the collections, the number of items harvested from each collection, and a description of the collection. Avano data providers include the Alfred Wegener Institute for Polar and Marine Research, the Aquatic Commons, and the Open Marine Archive.

The Scientific Commons (http://www.scientificcommons.org/) is a service provider which harvests a wide variety of scientific topics. The Scientific Commons has harvested 23 million records from 9 million authors in 950 archives. Searching a topic produces very large numbers of results (somewhat Google-like.) The search results can be filtered through only a limited number of options (Year, Language, and Sort Mode.) I think large-scale service providers such as the Scientific Commons call to question whether such services providers scale well for topic searching. The data providers to Scientific Commons are unidentified.

PLEIADI: Portal for Italian Literature in Open and Institutional Archives (http://www.openarchives.it/pleiadi/index.php?sel_lang=english) incorporates a variety of functionality in a multi-featured portal. Features include personal accounts, recent news, a wiki, RSS, and related web resources. Searching PLEIADI, it appears that many of its data providers are Italian university repositories.

Generally, I think it is most important for a service provider to harvest relevant data providers. It seems to me that service providers specializing in a topic can be more useful (especially searching on topics) than large-scale comprehensive service providers. In addition to harvesting relevant data providers, I think it is important that service providers have powerful search interfaces that provide users with sufficient filtering options and other search options.

EPrints

I thought it might be a good idea to try using the Library of Congress Subject Headings to describe my collection (a collection of evaluation reports in PDF) in EPrints. The first two levels of Library of Congress Subject Headings were insufficiently granular to distinguish individual collection items. In this sense, I think the problem with using LCSH with my EPrints collection is similar to the problem with LCSH in highly-specialized libraries. If I use EPrints again for this collection, I may create an alternate subjects taxonomy. A higher degree of granularity (distribution across more subjects) would make the subject browse feature more useful. Because my collection items are closely related in terms of subject, maintaining consistent subject terms has not been a problem. I used "uncontrolled keywords" to enhance the level of description. The uncontrolled keywords are not, however, by default browsable.

Drupal

I think Drupal should be a suitable content management system for my collection, but I probably need to work with Drupal more get the collection to where it needs to be. My collection consists of reports in PDF format which are best described with usual bibliographic categories (title, author, publisher, etc.) Primarily, I think the collection should be browsable by title, author, and keyword. I think Drupal should be able to do this, but I did not achieve this functionality within the time allotted.