Scott Leslie just pointed me to a Erik Duval's weblog which points to a couple of presentations from the Dublin Core 2004 Conference in Shanghai.
I'd missed these links when Erik first posted them (the "DC2004" title didn't grab me - doh!) but man, are they great presentations (well, the .ppt files are great anyway - I can only assume the presentations were better!)
They are basically talking about the changing role of metadata - how the html entry form is dead/useless for this. How multiple simultaneous metadata descriptions of resources are encouraged. Using multiple metadata schemas (LOM, MPEG, DC, whatever), as part of an active workflow rather than an ordained-from-above authoritative approach. About learning from cool new tools like Flickr, del.icio.us, gmail, etc...
It is excellent to see some high profile presentations on this - it's exactly where the learning object world needs to head (I've posted some thoughts on that myself).
When repositories (and metadata) fade into the background as infrastructure, and workflows and communities form to facilitate the creation, description, and sharing of the resources via said infrastructure, we will start to see some really fundamental shifts...
The basic theory is that the conventional "librarian" approach is flawed for a few reasons:
- it won't scale - limited number of librarians
- too expensive - authoritative metadata is expensive to generate
- too restrictive - shouldn't have to hire a team of librarians in order to share a free resource
- too inflexible - what if something doesn't fit in the LOM? Can I use EXIF instead? why isn't the content itself used/indexed? Can I describe something several times/ways for different purposes? In the Librarian world, someone "owns" the metadata - I can't change or add to it. I need to be able to do that.
Some of the stuff King has been working on for APOLLO over the last year will actually make this stuff possible for us, which is also amazing...