SFX (an acronym for Special Effects) is an exciting new Internet linking product. Originating with research done by Herbert Van de Sompel at Ghent University in Belgium, SFX is a context-sensitive reference linking solution that allows users to click on an article record in a database, and display a list of options available for that article, which may include the online full-text version of that article, among other resources.
SFX is owned by Ex Libris (USA), Inc., and the Caltech Library System has signed a contract to use it with as many of our bibliographic databases as we can, starting with Web of Science.
SFX extracts metadata from a given bibliographic citation in an SFX-aware database such as Web of Science, and passes it to an SFX server. There it is matched against a set of pre-defined relationships to other resources, and displays a list of services that are available for that article. For example, the window may display Get full-text from [the publisher], Check holdings in Caltech's Library Catalog, Requestdelivery via IBID, or Lookup as cited author in Web of Science.
The SFX database includes tables that define:
- Sources such as Web of Science, where a user would be starting her search
- Source Services, such as get full-text, find author, show local holdings, etc.
- Targets, resources such as full-text articles, where a user would want to go
- Target Services, such as display full-text, display author, etc.
- Object Portfolios, which define what Sources, Targets, and Services associated with a journal.
In order for a resource to be an SFX Source, it must be able to present an SFX button on each bibliographic citation record. This button activates an OpenURL (a standard format proposed by SFXs developers), which encodes SFX data (author, journal title, volume, etc.) in a way that the SFX server can understand. (It may also contain instructions for fetching additional information about the citation.) The SFX server parses the OpenURL and finds its set of related Target Services. URLs for each Target are built on the fly when a service is selected. Once the user makes a selection, a Perl program takes the metadata from the original citation and builds a URL that includes the resources domain and its standard link-to syntax.
Most information providers already have standard link-to syntax for their resources. Many include standard numbers (ISBNs and ISSNs), volume and issue numbers, starting pages, or author names. For example, one syntax may include the domain name and a unique journal identifier, while others include an abbreviated journal name followed by the journal year and issue number. Unfortunately, not all potential Targets have sufficient URL standardization to allow SFX to jump to a specific article, so the user may have to search through a table of contents, or even select a journal from a list of titles.