What follows began as a summary of my VALE symposium talk but I have been thinking about the talk and related items and, well, here goes. For those interested, I have been told the slides and videos will appear here shortly:
My talk touched on several themes but I want to elaborate on that part of the talk that discussed consortial software and particularly FulfILLment(tm) because there has been interest in it since VALE.
Evergreen’s OpenSRF is the service-oriented architecture that provides the structure that allows Evergreen to run a system on a laptop. It is the same architecture that allows Evergreen to scale across hundreds of servers to handle the large transaction loads of a distributed resource sharing network like PINES.
Consortia are a tool used by libraries to do many things as Mike’s earlier post outlines. There are consortia that are used for purchasing, those that are loose combinations for other purposes. Only a few historically have attempted to do what PINES has done: to combine their catalogs into one and run a large network with a statewide borrowers’ card. Union catalogs have been much more talked about than actually observed because they are difficult to do in the card catalog era and not much easier in the digital era until recently. And, of course, before Evergreen handling the transactions load on a large, distributed, and busy network was beyond available software. True, there are three or four (depending on final figures) US and one Canadian library that circulate more than PINES based on 2005 figures but from many fewer locations, fewer bibliographic items, fewer database edits from fewer catalogers, and not spread over a state, with all that entails. I also discussed the notion that there is a latent demand for software with such a capability in that post.
Our experience with such a resource sharing network has led us to conclude that:
Welcome to the long-tail, Google world. Our users can see a world without Balkanized information resources and they want it.
Of course, with the open source movement drawing attention, we find that the vendors of proprietary ILSs are more open than we are—just ask them. Similarly, everyone does resource sharing now from individual libraries which do not have a single union catalog. Ask them how many instances of their software they run at the consortium and have them describe how they migrate records from one system into the union catalog.
Be that as it may, how did we come to this place where we have all these individual libraries doing their thing? The pattern is an old one and the term of art to describe it is “information silos.” I defined information silos for the talk as separate and barely communicating collections of information. Interlibrary loan is a sluggish, time consuming, and awkward way of sharing materials between libraries. Its cost in time and complication reduces its use and its effectiveness. ILL is an example of bad communication between libraries where separate libraries barely make materials available to other libraries.
PINES breaks information barriers between member libraries and their users and removes the awkwardness of ILL by making it a type of internal circulation transaction, termed a “hold.” Not a kludge bolted on between library silos but an integral part of each library’s system and accessible to a library’s users as they search the catalog. Holds are much easier than ILL so we find users availing themselves of the option in astonishing numbers. You are seeing not only the long tail but that part of the tail that is just slightly longer than an individual library user’s collection—the slightly longer tail, if you will. Why do we have silos? In this talk, I offered these possible explanations:
There may be others. For instance, I fear in my darker moments that some of our colleagues do not like to share but my sunny disposition will not countenance that thought at this time.
One of the questions that folks ask us is: aren’t you just copying the legacy vendors? It looks that way, doesn’t it? Because in this market, librarians expect “integrated” systems. That is how Evergreen presents. But, underneath, it is completely different. It is modular and built on the OpenSRF architecture I mentioned.
Of course, everyone is modular these days—just ask them. We mean modular in the sense of distributed nodes in clusters not merely modular as a means of monetizing components. An example: If your Evergreen installation outgrows its servers, you buy another commodity server, install the software, and the new box and old boxes start talking and away they go. You don’t have to go out and by another million dollar server because the software you are using was designed when mainframes were in vogue and that is how they did things way back then. The future will give librarians a host of different solutions that we can only speculate about. One we discussed at VALE was that if you didn’t like the Evergreen’s circulation module, you could replace it with one that you liked better. Mix and match of independent modules seems very likely will be a part of the library software world of the future.
However, we can do a bit more than speculate now about one development. Mike Rylander, VP for Research and Design, and I gave a talk on January 17 at the State Library of Ohio in response to a Request for Information about upgrading a resource sharing model they have used, MORE. Our proposal was called FulfILLment and Mike’s earlier post has the diagram and more discussion of it. It was an attempt to get past the silos in another way. This one uses the proven Evergreen backend for real-time processing with opportunistic connectors to legacy vendor software. “Opportunistic” because many things that we would hope would work don’t.
A user of a FulfILLment(tm) library would have a better experience now because FulfILLment would simulate much of the capability of Evergreen to make a virtual catalog available at the time of search. Would it be as good as Evergreen? No, but it would be pretty close—certainly better than what most libraries have now.
FulfILLment is Mike’s idea and his thinking is now beyond where he was in January and we are seeking development money to build it. After it was developed, it would be released under an appropriate open source license for use by the community.
The Ohio State Library was facing a dramatic budget shortfall and that is where the project stood until VALE. Now, we seem to have more interest. I certainly hope so. If libraries are to build or to have a part in the information environment of the 21st century, we better get going real quick, now. I submit that if the ARL libraries, for instance, were reachable through a FulfILLment(tm) interface, it would have a profound effect on scholarship and the learning environment at ARL member institutions—as it would at any set of academic libraries using FulfILLment.
And why not public libraries? If all or a substantial number of libraries in PALINET or SOLINET were reachable through that kind of interface, it would also have a profound effect on the use of member libraries and be a big step to moving libraries closer to the center of the information society instead of out where the buses don’t run.