avoiding IR failure...

Reading and processing Dorothea Salo's valuable preprint on institutional repositories and their problems. Several thoughts come to mind....

First, the fact that the article exists really emphasizes the need, in my view, for librarians to have faculty status. Like it or not, librarians need to publish in controversial areas such as scholarly communication and on intellectual property issues. As Salo notes regarding research on IRs, "doubts about the viability of institutional repositories have been kept quiet or denied altogether," which I attribute in part to the decision of many larger ARLs to make their librarians part of the support staff. At Wisconsin, where Salo works, librarians are academic staff, but I still believe that my point is valid. I have to admit that my own thinking has shifted on this quite a bit. I was more supportive of the librarians-as-professional staff concept following my stint at NLM. But for publications in the areas I cited above, librarians need the same protections as other faculty.

Key point made by Salo: The best selling points for an IR - digital preservation, preserving access, and increasing the exposure of research - are not that relevant to university research faculty, for reasons described by the author.

Throughout the article, Salo discusses the capabilities (or lack thereof) of the DSpace repository software stack. As one example, she notes the inability to add text containing hyperlinks to Dublin Core metadata; this is true as of DSpace 1.4, the current, non-beta, version. Personally, I find the inability to format metadata (additional example: italicizing fonts as needed) to be a major drawback for DSpace. A bigger problem is the lack of an easy-to-use API to support local customization of DSpace. In doing so, she touches on one of the weaknesses of overly complex OSS solutions: "even though it is possible and legal to modify the code for local needs...doing so is deeply unwise from a sysadmin's perspective."