More on classification

When I asked Where's the fight-back from formal classificationists?, Matt pointed me to some very interesting websites, including two at LibraryThing. These projects are great, but they're about bridging the gap between formal classification and folksonomies, rather than shoring up the somewhat dilapidated current state of formal classification.

What I had in mind was a program such as this:

Imagine a theoretical classification scheme, which is still widely used in libraries and would like to be widely used on the Internet. The guiding hands decide that if they are to survive on the Internet they must adapt, to this end they:

1. Publish a series of elearning modules, to enable a wide variety of users to upgrade their skills with respect to the scheme.
   1. first-year students needing to find the recommended reading material for their courses
   2. research students looking for peer-review papers
   3. librarians looking for professional development
   4. software developers looking to use the scheme in their application (more on the later)
   5. content creators looking to classify content
   6. content consumers looking to classify content
   Only the last two classes of user are genuinely new, of course, so much existing content could be migrated initially from the current systems.
2. Publish the results of these courses in machine readable formats, so users could link back to them.
3. Change the copyright licence on the text of the scheme to enable third parties to use the materials in information systems.
4. Work with content creation, storage and distribution systems du jour, to enable them to use the scheme correctly. This is most likely include:
   1. encouraging users to learning how to use the scheme correctly (via the courses outlined above)
   2. encouraging systems to link to course results, so the automatic systems can check that classifiers have been sufficiently trained.
   3. encouraging communities to develop traditions of quality classification. One could imagine a scheme such as slashdot's meta-moderation working very well.
   4. linking between systems in meaningful ways.
   Technically these could be achieved in a number of ways, including direct participation in the development of the software (particularly easy in the case of open source software), development of libraries to do these things in the languages used by such systems (currently mainly Java, Ruby and PHP).
5. Focus staff development resources (including elearning modules discussed above) to enable librarians in bricks-and-mortar libraries to meaningfully and usefully answer user questions about finding materials on the Internet by leveraging the fact that the classification scheme which they already have person-years of training with now covers a much larger field.

There are a number of people and groups who see themselves as generating content for the ages (think Project Gutenberg, Wikipedia and some of the better photographers on flickr) and I believe many of these would more than welcome systematic classification scheme. The existence of should a scheme would also solve problems such as representation of foaf:interest in a global context and other interoperability issues which arise in the web 2.0 world when trying to describe real-world concepts without dependence on a particular system.

If such a system did take root successfully, it would be very hard to supplant it, because just as a second classification scheme is of very limited use in a library, a second classification scheme would be of very limited use on the Internet.