Amazon recently released its S3 service - a network storage system.
Amazon S3 provides a simple web services interface that can be used to store and retrieve any amount of data, at any time, from anywhere on the web. [Amazon.com Amazon Web Services Store: Amazon S3 / Amazon Web Services]It received quite wide coverage. One interesting feature was the absence of a feature - the user interface. It did not have its own user interface: it is available only through machine interfaces. One of these is BitTorrent. So it is built from the start as a network service, a service that other applications communicate with.
In these pages, I have wondered what it would be like only to deliver library services through common services (Flickr, delicious, Technorati, ....) and browser tools (toolbars, bookmarklets, ..) ? The S3 announcment prompts another thought experiment along these lines. What would it be like to design library services with no user interfaces at all, with no direct web presence? To design them in such a way that they could only be consumed by other applications. These application might be prefabricated workflow managers, like course management systems, or the various tools which which we are self-assembling our work and learning environments, such as RSS aggregators, my.yahoo, and so on.
What this would force you to do is to think about the variety of places that it would be useful to create a library presence and then to think about what shape it would need to take to be consumed there. So, people use RSS feeds in a variety of places: what would it be useful to expose as feeds? Stepping back a little, it would make you think about what should be URL-addressable. If you wanted to create reading lists in Deli.cio.us, for example, individual records would need to be URL-addressable. So, even before one began to think about web services, APIs, OpenSearch, and so on, there are some interesting questions to ask.
I am not at all suggesting that we not expose services through user interfaces: this would not make sense. I am suggesting that imagining that we have no user interfaces would make us work through some interesting questions about where and how people would benefit from being able to interact with library resources.