RefDB and Open Source Development
For an academic in need of reference management, storing and retrieving references only covers one need. The other significant need is document formatting. This is a difficult task, as journals and publishers all have their individual styles.
AFAIK the only project to adequately deal with formatting bibliographic citations in XML documents is RefDB. For a long time I’ve been of the opinion that open source projects should adopt RefDB as their core storage and formatting engine, and focus on UI and other innovations.
RefDB is based on open source relational database storage, but has an abstraction layer that allows support for MySQL, PostgreSQL, and SQLite. It has a clean client-server design, and includes a Perl module that communicates with RefDB via its daemon; designed precisely to interface with a GUI. Apparently writing such modules – in PHP, or Java, or Objective-C – is trivial; roughly a day’s work. Moreover, the developer – Markus Hoenicka – likely knows more about formatting XML-based bibliographies and citations than anyone on the planet.
For whatever reasons, my argument has been generally rejected.
I have thus come to the conclusion that it might be a better approach to modularize the pieces of a successful bibliographic system. While RefDB has a functional web interface, it lacks some of the functionality I’d like to see:
- Support for rich annotation, with xslt rendering. Beyond rendering for the web, I’d like to click a button and have a pop-up with configurable transform: to LaTeX, DocBook, TEI … whatever.
- Hot-linked queries on author names, keywords, etc.
- Better record entry UI, which is currently just a text field in which you can paste RIS or BibTeX data. Ideal would be a full-blown configurable form UI modeled on MODS, but I could imagine also something more modest; perhaps a selector that would insert the proper XML skeleton based on the reference type?
I’d really like to see this implemented using the xslt-based approach of Syncato/Cocoon/Popoon, which could easily be adapted to different storage systems. So, some might use it with RefDB, others with a native XML DB solution, and still others with flat files.
Any takers? Alas, beyond basic xsl, I can’t code!
Great idea. Along similar lines I had been thinking about a standard XML representation of subject guides to make content in the ResearchGuide project interoperable (see http://researchguide.sf.net/).
I have some PHP skills but I don’t see myself having time for coding up the kind of thing you’re talking about. But I’d be interested in hearing of any developments.
Looking at the interface, it seems the data model is pretty simple, so it shouldn’t be hard to design an xml representation of it.
I wonder, is there some way to think about the metadata associated with these guides in a broader way that might link it to other kinds of applications? How, for example, do you see these related to the subjects and topics listings in MARC/MODS? Dublin Core? The learning module stuff that I know virtually nothing about?