Yesterday I (and several other library folks) had an interesting meeting with the art museum people on campus. They were kind enough to give us a tour of the art museum’s new content management system. I was particularly interested in seeing this system because I haven’t had much exposure to the museum world (but have been curious how well XOBIS would fit with their requirements — this is not why I was there, of course, but was a personal agenda of mine).

The first screen we saw was a screen of buttons with things like: Object, Event, Exhibition, etc. One could create a record for each of these by clicking on it (you then get a detailed editor for that type of record). There were about ten or twelve of these record types (I didn’t actually count them) and though they didn’t completely map to XOBIS’ ten principle elements there were some interesting overlaps. For instance, Object and Event would correspond to XOBIS principle elements, but Exhibition would be a type of Event in XOBIS.

We were told that, while the system had predefined databases tables for the specified components of a record, the label used in the display was completely configurable (this was true at the editor level; I don’t think it was true at the button screen level). I see no reason why the same couldn’t be done for XOBIS. An editor might say Exhibition, but when it stores the record an Event record with relationships to the types of things that an Exhibition type of Event might have would be stored. There didn’t look like much in the system that couldn’t be handled this way.

One neat thing about the system was it seems to allow a much richer range of possible connections between these record types (than most library systems). We were unable to find the relationship page, though, so I couldn’t see how much information was stored about the relationship itself. Also, the administrative information that was stored about an object did seem a little brief. It looked like only one set of rights management info could be recorded, but what if an object or reproduction may be used one way for photocopying and another for display on the Web?

Overall, I was impressed with the system. I’m not sure it would work for a library (authority control was there, but it seemed a little limited — you either had to import records (which, for us, would mean a split with the catalog) or connect to particular types of external resources (that was my understanding at least)), but it looked like it was exactly what they needed at the museum. One person at the meeting talked about how they had been using an offline system so being able to move online was a great advantage. I’m surprised the art students haven’t requested an online system long ago.

One thing I didn’t like about the system, though, was that to get a web interface or XML out, one basically has to do a system dump (or snapshot). There is no live interaction with the database outside of the staff editor. This makes working with the database a little tricker if you want it to share info with others, but at least system does output XML (using XSLT to generate the web view). It would be even better if it put out VRA, but there are things in the system that would need a higher level wrapping schema too (and writing the XSLT to get VRA shouldn’t be too difficult).

The meeting was an interesting one. They expressed interest in the work that we were doing so hopefully we will be able to have some more meetings and share some ideas. I’d definitely be interested in seeing more of the system at work once they bring it up and are using it in production. It would also be interesting to hear about their experiences migrating from what they have to this new system.