Campus Technology has a great interview with Michael Keller of Stanford University Libraries on archival digital librarianship - how libraries can and must preserve the past by both creating digital copies of historical documents and by maintaining born-digital documents.
It's that second category that has a lot of really interesting possibilities. I wonder if one possibility for outreach by academic libraries is to graduate students and faculty on how to back-up and archive their own work, possibly in conjunction with IT departments. I know, for example, that my own dissertation data is in data analysis program format that's already one release behind.
And it's data we should be worried about archiving and releasing. While faculty may have mixed reactions about releasing their articles as open access, there are strong, important scientific norms about providing access to data sets, so it's relatively easy to start with them - but it's also more important. Anyone can just save their article as a .txt file and preserve the sense of it in a universally readable format. (Math-heavy articles are an exception, but those scholars have largely made the move to open source LaTex, which I expect will be accessible for years to come.) Datasets, however, need special handling to be usable into the future, and the files can get much, much larger, making them challenging. They seem like a good starting place for institutional repositories when faculty are reluctant.
30 minutes ago