I don’t think libraries should give up on providing valuable ‘discovery’ services. But our “unique value proposition” is probably more about delivery — nobody else is going to do for free what we will (‘free’ in the sense that our business model already involves providing such to our patrons). We can probably get a better ‘bang for the buck’ by focusing on that aspect of our services more, taking what we already do and making it as incredibly convenient and user-focused as possible.
What if… you could send a txt message from your cell phone with an ISBN, or better yet an MMS from your cell phone with a photo taken of an ISBN barcode, send it to the library, and the library would automatically request that book by the appropriate delivery mechanism for you? It’s in the library, okay place a request/hold through the ILS, it’ll be waiting for your at the circ desk (or delivered to your campus mailbox if you’re faculty at my library). Not at the library? Okay, ILL request placed. A confirmation txt is sent back to you saying what happened. (Or an error if the ISBN can’t be recognized, I guess).
That’s be really really cool, huh? It’s totally within the realm of possibility. Umlaut (you will be unsurprised to hear me claim) is a great platform to build such a service on. If I actually had time to dedicate to it (which I don’t), I think I could have something working in a few weeks or a couple months. (If I were in charge of staff allocation, which I’m not, this is what I’d be telling me to work on, heh.)
The biggest barrier would be that our actual document delivery/request services lack any kind of API. Some kind of hacky screen-scraping might be required, and when you add in the need to authenticate, it would be really hacky screen-scraping.
OCLC, an API to ILLiad please? Pretty please? The API should allow a trusted system to place requests on behalf of a specified user, and return the ILLiad internal request ID; cancel a request with a given ID on behalf of a user; and list all outstanding requests and their statuses for a given user (Atom publishing, or an extension thereof, might be a good format for the latter, if not for all of those.) I think the DLF ILS model API might also provide some suggestions for some of those actions. I could probably hack up an API to do that on top of my ILS, since at this point I’m pretty familiar with the inner workings of my ILS and have direct db access to it’s underlying fairly well structured (one of the plusses of the cancelled Horizon) rdbms. But ILLiad, phew, I don’t know.