I recently visited the ELib Open Journal Project at Southampton University (see <URL:http://journals.ecs.soton.ac.uk/>). My hosts were Leslie Carr (the project manager) and Steve Hitchcock (one of the OJ researchers). The Open Journal project is part of the ELib electronic journals strand, but they have developed technology which may be of use to some of the Access to Network Resources projects. This trip report details some of the work that they showed me and also outlines some possible ways in which the Open Journal Project, ROADS and the other ANR services may be able to cooperate.
The basic idea of the Open Journal project is to explore the concept of "opening" up the traditional journal model when it is moved into an electronic environment. They are attempting to do this by providing technology to basically allow hypertext links to be retrospectively added to existing "closed" documents. These hyperlinks are derived from specialised databases known as linkbases. Communication with these linkbases is mediated by the Distributed Link Service (see <URL: http://wwwcosm.ecs.soton.ac.uk/dls/dls.html>).
The hyperlinks in the linkbases are related to words or phrases that appear in the original documents. A linkbase reference might apply to a specific word in a specific place in a specific document, any instance of a word in a specific document or any instance of a specific word in an arbitrary document.
Originally the user interface to these was supplied in a number of ways. Firstly they developed a tool for MS Windows, Mac and X11 that allowed any highlighted section of text to be compared to the available linkbases to see if there are any related hyperlinks. If there are, the user could select one and the tool will communicate with an instance of Netscape, or any other web browser, to download the object that the hyperlink points to. This is quite neat as it means that you can look stuff up in linkbases from inside any Windows, Mac or X11 application that permits a selection to be made. Could be handy if you are sitting in Emacs or your favourite wordprocessor writing a report and want to see if there are any references in the linkbases to a specific product, project, phrase or whatever.
However now the project is aiming for operating system independence by concentrating its effort providing access to the linkbases via a special proxy server that can add in extra hyperlinks based on the options that the user has selected. For example the user might want to only see hyperlinks specific to that document or only general purpose ones. The user communicates his or her preferences to the proxy server using either a small application that sits alongside his/her web browser, or by using controls embedded in the HTML document itself.
Both Les and Steve were keen to stress that they viewed their software as providing increased access to network resources as well as allowing existing journals to have neat navigation easily added to them. Although the Open Journal project is using the software to integrate existing journal articles into the web's hypermedia, there appear to be plenty of other uses for it. They told me that they wanted to form links with ANR services if possible as they could see a lot of "synergy" between what they are doing and what we are doing.
One suggestion they had was that ANR services might wish to generate linkbases for their subject specialisms which their users could then access with the Open Journal project's interface. For example, SOSIG (say) could set up a linkbase using its keywords from the templates it holds in its ROADS database, with entries pointing either directly to the resource or possibly to the tempbyhand.pl script on the SOSIG machine. Social scientists could then point at SOSIG's linkbase using the Open Journal project interface and use it to access quality network resources after selecting words in their Word documents or whilst reading other HTML documents.
There may be advantages for both the ANR SBIG and the Open Journal Project in doing this. The SBIG would be getting kudos for basically reusing their existing existing database in a new form and providing their community with a handy new service. The Open Journal project similarly would get some kudos for having their link service widely used and also ensures that subject specialists are generating metadata for their linkbases. We chatted about this idea a little and I promised that I'd suggest it to the ROADS using ANR services (so here it is!).
The Open Journal project's software is definitely worth a look at and I can see a number of other uses besides the original purpose. For example, it appears to be a handy way of dealing with lots of hyperlinks that appear in lots of documents from the same server. Handy if you're looking after a large number of documents and somebody goes and changes one of the URLs that many of the documents contain a hyperlink to. Instead of having to edit each document (or write a Perl or sed script to do it for you), you could just tweak a generic entry in the linkbase for the word or phrase that the hyperlink is usually described with. Or how about using it to act as a sort of global annotations server where a suitably tweaked HTTP caching proxy queries a distributed set of linkbases for annotations to add to each document that passes through.
Anyway, that's my report folks (which hopefully justifies my 7.50 expenses claim. :-)). I'd just like to thank Les and Steve for putting up with me for a day, feeding me a yummy veggie pasta meal and letting me read my email whilst I was there, as well as showing me their current linkbase developments.