There is already a collection of notes and links on the web about OR2007. Some blog posts summarizing OR2007 presentations can be found using the technorati tag "icor2007". And links to many presentations, demos, and attendee photo collections can be found with the del.icio.us tag "or2007". So I'm going to continue my own retrospective on the conference.
One good aspect of this conference was the number of presentations of individual and institutional experiences developing, implementing, and managing open, institutional repositories of publications, reports, and data. This is a dedicated community, and a good deal of progress is being made building successful repositories, and the presentations reflected this.
A case study about the American Geophysical Union (AGU) digital archive suffered a total presentation interoperability meltdown (at least that's what it looked like to me) and the presenter was asked to just speak without notes. So he told the story of the AGU's migration from a print-based publication to one that will be completely electronic by 2009. After that the electronic version will be the version of record. The presentation made clear that substantial effort is required to move from a paper publication to one that is electronic only. One example is illustrative of the kinds of problems every institution will face if and when they undertake this move. When the journal was a paper production there was little interest in the names of the files used to create the print galleys. The goal was getting the print copy done on time, and once that was done, there was little interest in organizing and maintaining the source material. The distributed print copies represented the copy of record and the archive. But when the journal started to publish online, the lack of standards for file naming was a problem. It seems like a small detail, and in retrospect it is no surprise that collections of idiosyncratically named items are hard to use and manage. And for those experienced in the constraints of computer-based file systems it's hard to think that file names wouldn't be regularized from the outset. But everything new yields surprises and the stories of surprises contain important information for others who are, or will be, starting similar projects. What surprises one person can be a helpful hint for someone else, so these kinds of stories should be encouraged. I know that some repository folk provide these stories on weblogs and, no doubt, in e-mails. But more is better, and the more frankly we talk about our lived experiences with building and deploying open repositories, the more quickly we can create sustainable work practices for new institutional repository managers and users.
Perhaps at other OR conferences there might be a session devoted to stories of the day-to-day work of running the repository -- repository operations. I don't know how many would volunteer stories, but it would be great to hear how people make it work.