I can only assume that the only reason someone didn’t point this paper out to me during my recent thrashing about concerning the difference between ‘institutional’ repositories and ‘learning object’ repositories is that, like me, they had never seen it before (or maybe you’re all just sadists and like to watch me flail about in public!)
Well in any case, hallelujah! This draft paper by Neil McLean and Clifford Lynch from June 28, 2003 is in my mind a model of clarity on the reasons for why these beasts are different (for one, the ‘transient’ versus ‘archival’ nature of their contents) but also why and how they need to interoperate.
Which is where I’ve landed on this topic – we need distinct types of repository software because they fill distinct end-user needs. But by implementing both common open protocols and using structured markup languages that can be mapped, we keep open the possibility of interoperating if and when this make sense. And I stress that last ‘if’ – the next piece in the puzzle I am waiting to see are convincing use cases, or even better yet convincing demonstrations, of search interfaces across catalogues of heterogeneous materials (e.g. records for books, ‘eprints’ and learning objects all at once) that don’t just confuse the matter entirely. – SWL