Today was a busy day. Over lunch, Purdue Librarian Mike Witt regaled me with historic tales of the early days of Purdue's institutional repository (IR). He should know: he started and spearheaded it. He's also the CS librarian and a huge open access advocate, which explains another reason why we were sitting across from each other in the dormitory dining hall, chattering busily between mouthfuls of strawberry cheesecake. (Purdue visitors: Earhart Hall has an incredibly impressive repast. I would go again. Recommended!)
"The point of an institutional repository's collection policy is so that you can say no."
It's not about what you will collect, it's about what you won't collect - how you don't want to use your limited resources. Lunch with Mike was a joy; his casual comments (because he's deeply embedded in his field) expanded to a wealth of information when I tracked them down. For instance, the Panton Principles (rationale for open data) - how did I not see this before?
I learned a lot about how someone thinks about an institutional repository. (I know it's one view out of many possible ones, but still - getting a rich slice of that one view was extremely useful.) Content in Purdue's IR is organized the same way its departments are organized; interdisciplinary research usually comes out of Discovery Park and goes in its corner of the repository. There are common actions that an IR's workflow should handle well: consuming (downloading someone else's papers), submitting (sending in your own), accepting (what it sounds like), and batching (creating collections of papers). Make sure you check and know all four.
Mike confirmed some other things I had been thinking about, like how catering to faculty self-interest ("ah, I can track publication downloads and use this for my tenure case!") actually works well. Mandating something that should ostensibly be in faculty's self-interest ("only publications listed in the IR count for your tenure evaluation!") is... hard, and I'm not sure if I agree with that goal (I still live in a dream world where people do good things without being required to do them, apparently) but it's got an undeniable impact on compliance rates.
On the way back, he introduced me to the work of Jason Priem, who looks at altmetrics - can we more accurately measure scholarly impact if we look at social media instead of merely journal citations? - and whose open access koans made me grin. Apparently Jason is speaking at Purdue in 2 weeks - I'm totally going. And apparently Jason is getting his PhD at the same place as Bryan Behrenshausen, last summer's opensource.com intern (and the man behind the POSSE profiles from last summer - we're currently having a debate about the future of POSSE, by the way).
My brain is swimming -- and I haven't even started my German translation homework yet. It's been a good, good day. Extremely good. A day full of The Learn.