I recently answered a kind inquiry about the lessons learned during the 3-yr fixed term SALT Project at Stanford. I thought my informal answers were worth posting here.
1) The foundation of
SALT, the database, was very similar to DSpace. When I arrived on the project,
SALT was using an in-house developed database that the Stanford Libraries had created before DSpace became popular. After the completion of our
SALT Prototype, the
SALT database is being ported to Fedora. So open-source tools are viewed as very desirable. What made
SALT unique and different was the user "access" side: the ability to enable and deliver visually-oriented views into the material. We showed information summaries in "graph network" maps and timelines, and then the user could drill down into a faceted browser to search for individual objects (papers) of interest. We used Berkeley's Flamenco for the faceted browser, later replaced by UVAs Blacklight. But the visual representations of the faculty legacy was the thing that was the most unique and drew the faculty's interest.
2) The
SALT Project was a collaboration between the Stanford Libraries and a professor emeritus of Computer Science. So the project had a flavor of research, even though the intent was not to develop a lot of software in-house but to package open source solutions to achieve the documented requirements. The team was asked to apply for an NSF grant, which we did, so the objectives had a strong flavor of research. As you can imagine, the requirements were fairly visionary and it was a challenge to deliver working examples of the elements. However, we mostly tried to find existing solutions as I mentioned above ... and for each solution chosen we evaluated several. We implemented the
SALT Prototype in 3 iterations. With each prototype release, we got a better understanding of the goals and requirements of the various stakeholders (faculty, archivists, engineers, curators, researchers, management). In our final version, all the major software functions were off-the-shelf and the software we developed just connected the elements of the tool chain together.
3) The faculty reaction has been very positive. I interviewed a handful before and after delivering our prototype, and they were very interested in documenting their legacy and especially tracking the flow of their teaching through their students. The "graph network" map was the key to drawing them in to offer "the backstory" and new knowledge. They were most willing to offer information verbally, rather than typing (the older faculty were challenged by eyesight and typing). They particularly valued that the system was largely automatic, because they are very limited on time to enter data or insights.
4)
SALT is moving forward as a "production" project within the Stanford Digital Library group, mostly taking steps to integrate the object-level documents into the library's
standard catalog. I do indeed anticipate some problems still to be solved, for example with the dis-ambiguation of names and tag metadata. There is not a good point-solution for this problem, but Stanford's archival staff were looking to the Library of Congress classifications. There needs to be more thought and integration with traditional archival data-views, Stanford's archival staff and the researchers using the collections were more comfortable with finding aids in EAD format; I think these could be easily generated as an alternate view. Ultimately the biggest problem is the volume of faculty archival information far exceeding the archival staff's ability to process it, so all stakeholders will need become more comfortable with automated solutions. Lastly, there is an unsolved problem with copyright and privacy liability vs offering access, especially with regard to email. Email is becoming more and more a way for archiving and collaboration. See also: papers by Cathy Marshall of Microsoft.
I continue to think that archiving of faculty papers is a very important area, as faculty papers are more and more born-digital an faculty are finding solutions for self-archiving. I found that the cutting edge profs were already getting their staffs to digitize important papers and lab notebooks, and create online archives with self-description. I think it will serve the University and patrons to help the faculty be successful with these efforts, and in the process to organize their material in a way that is more easily ingested into the Library's preservation system.
