Portrait of a vision unfulfilled
It was developed, I beleive, before there really were any realistic commerical ERM systems. (Some would suggest that’s still the case). For reasons I don’t entirely understand since they were before my time, it never really caught on as intended.
It was intended to support back-end workflow (ordering, tracking, etc., of licensed electronic resources). And I know that the developers spent a lot of time working with users here in a “user centered design” process to try and ensure it did this appropriately. But for some reason, after it went into production, our local staff was never interested in using it as such, and never did. I guess it didn’t end up meeting their needs after all.
Hermes was intended for use across our various libraries, but really only had much adoption at all at the general purpose MSE library.
The developers originally hoped that this software would be used by other institutions too in a collaborative open source project. That didn’t end up happening. I think that most other libraries weren’t quite ready for this yet — interestingly, over the past few months I’ve received several inquiries about Hermes, but had to tell people that it was pretty much a moribund project they didn’t want. One of the developers, Nathan, also left our institution soon after the Hermes roll out, which of course didn’t help.
It was written in ColdFusion, which made sense at the time I’m sure. But we ended up with a software package that none of our technical staff were really that familiar with, and could not support very well. It’s development basically stopped after Nathan left (there was more than a one year period in between Nathan’s departure and my arrival; by the time I arrived, it had already been de-prioritized and put onto life support only).
What Hermes was still being used for was a public discovery system. A searchable directory of licensed (and some free/public) electronic resources (and some non-electronic too; it ended up being rather a mish-mash of a directory) here at our institution. With categorization by subject.
When I was looking at our web site one day, and realized we had a) one link that you could click on to get a list of subjects, and then click on a subject to get a list of databases (Hermes), and b) another link you could click on to get a (somewhat different) list of subjects and then click on a subject to get a list of databases, this time with cross-searching ability (Metalib, our federated search product)…
…I realized this was confusing, did not serve the users well, and we could combine our ‘database directory’ patron service with our ‘federated search’ patron service. Making a better more consistent less confusing user interface, and at the same time getting rid of one more back-end database with redundant information that needed to be manually maintained (how many different systems do we have that need to know about all our licensed databases? One fewer if we could get rid of Hermes).
Xerxes, a Metalib front-end developed by David Walker at California State University and already in use by several institutions, was the answer. In order to make Xerxes do everything we had been relying on Hermes for, I needed to add a couple features: automatic generation of a naive EZProxy config file from the Xerxes/Metalib database; embed/passthrough functionality in Xerxes for subject pages and individual databases.
We deployed Xerxes a few months ago, and after a suitably long and painful process of trying to get rid of any remaining links to Hermes, today we turned off Hermes.
Of Risk, Success, and Failure
So was Hermes a failure? Well, I don’t think it accomplished most of what it was intended to. What it did accomplish was probably not ultimately worth the development resources that went into it. I guess I’d call it a failure, but I can be a perfectionist sometimes. And to clarify, this is no fault the developers, some things were just out of their control.
On the other hand, it held us through until we had something better. It provided some patron web services that we had no better alternative for providing — until we did. It kept working, even after the original developer left, staying up and serving our users, until it was no longer needed.
Innovation entails calculated risk. Sometimes you aren’t going to accomplish what you intended, and this is inevitable if you are going to try to innovate and improve services for our patrons. If this is the biggest kind of failure you ever have, you’re probably lucky. We should be prepared for even bigger failures than this if we really want to innovate, although if we’re careful and lucky we’ll never encounter one.
The New King
(I feel like I’m in a Greek drama now).