My own experience: I was involved (in Cambridge, England) in a project across a "federation" of secondary schools, all using this one database of learning resources. Teachers could log on to the system, enter the
URL of a learning resource, then type free-form text "keywords" that could be searched through to let other people find the resource. This didn't work at all. For a start, some teachers didn't know what we meant by "a URL", and couldn't understand that they needed to
upload files to a webserver before they could be made into "resources". Then, they had to log in to a web-based system to enter the URL of their resource, so they needed to spend five minutes logging on and typing or cut-and-pasting their URL. After this, they had to enter some keywords - most either skipped this bit, or typed a full-text description, or typed "stuff", or used different keywords from others...
This proved to be a useful system because it gave me a good idea of how not to construct a database of learning resources. I would say that a good learning resource database:
- Should be quick and easy to use. The quickest and easiest method of getting a URL into a database that I could think of was to have a web page that takes the referer URL as the URL to be added to the database (don't ask for login), then asks for keywords to associate with it. Then, simply hook this page up to a button on the user's browser toolbar - any time they want to add a URL they simply hit this button, they don't even need to know what a URL is.
- Should use defined tags, clearly laid out in a taxonomy, preferably one that lots of other people use (e.g. the UK's National Curriculum) - learning resource creation companies are starting to release their products ready-tagged with metadata. Exactly how you attach each resource to appropriate tags is a good question. It takes a bit of explaining and concentration to get teachers to understand that it's important that resources be associated properly with metadata - it's hardly an interesting subject, and teachers have other, much more imporant things to think about. I thought that one idea might be to have the add-some-keywords form ask for a free-text list of keywords then, on submition, have these keywords checked against the ones in the database. Any that don't match up exactly and unambigously get a "Did you mean..." message next to them. You could also swap synonims for approved keywords. If the user is sure that the unrecognised keyword they've typed in is appropriate, then they could choose to have that keyword added to the database.
- Needs to be large-scale. Google and Flickr work because they have such large databases (obtained through much investement in masses of hardware and top-secret search algorythms) that even if you don't see every item, you've still got a good selection. In anything smaller scale, each entry to the database needs to be checked to make sure it is associated with appropriate keywords, has an appropriate title. Some resources expire after a while, or their URL moves. New keywords need to be added to the appripriate place in the taxonomy. All this is the kind of thing best managed by a professional database manager - the project I was involved with tried to delegate this stuff to teachers, but this just mostly annoyed them. Also, we only had five schools contributing to the project, and we started off with an empty database. No one wants to put work into an empty-looking database, so the database doesn't get full, so no one wants to...
I'm guessing that this is the idea behind the Moodle Community
Server - share metadata and reources aound organisations (the country? the globe?). Maybe even have someone sitting full-time in Moodle headquaters in Australia checking over the global database of resources. The only problem is how to pay for their wages - charge to subscribe to the database? Advertise company's products, Google-style, in appropriate places?
--
David Hicks