I am at the ELUNA (Ex-Libris user’s group) conference, and just saw a presentation on Ex Libris strategy from Oren Beit-Arie, chief strategy officer of Ex Libris, and Catherine [someone], URM project manager.
I was quite impressed. The URM strategy is basically Ex Libris’ vision for a new kind of ILS (post-ILS ILS?). [I talked about this before after last year’s ELUNA] I hope they make the slides and such from this presentation public, because I can’t quite cover what they did. They showed basically a structural diagram of the software (fairly consistent with what I wrote on future directions of library systems), and some mock-ups of staff user interfaces for workflow.
But my basic summary is: Yes! This is indeed a vision for an ILS that makes sense. They get it. I imagine most code4libbers if they saw this presentation would agree, wow, yeah, that’s the way an ILS should actually work: supporting staff workflow in an integrated way that actually makes sense, modular and componentized, full of APIs and opportunities for customer-written plugins, talking to various third-party software (from vendors (of various classes) to ERP software etc.), etc etc.
The vision was good. What’s the execution going to be? So enough praise, let’s move on to questions, concerns, and some interesting implications.
The timelines given were pretty ambitious. Did they really mean to say that those mock-ups of staff interfaces Catherine showed are planned to be reality by the end of 2010? The software is ambitious enough to begin with, but on top of that the mock-ups shown were heavily dependent on being able to talk to external software via web services, and getting other people’s software that may or may not have those services available to do what’s needed in that time too… I’m a bit skeptical. [And with the staffing required to pull this off and based on pricing for other products… I would have to predict that end pricing is going to be in the mid six figures].
On top of that, when Oren supplied his timeline there seemed to be a bit of slight of hand going on that confused me a bit. He said that next version of SFX was expected end of 2009, and with that lit up a bunch of boxes on his structural diagram that he said this release of SFX would fulfill. If SFX is really going to fill those boxes for the integrated and modular architectural vision he outlined (with rationalized workflow and data management not based on the existing borders of silos that exist for historical reasons, and which SFX definitely exhibits—and SFX is not known for a staff interface that supports rational workflow)…
….then either SFX is going to become quite different software than it is now (by end of 2009?)—-or the execution is going to be significantly less than the vision offered.
Network-level metadata control?
Part of the vision involved a network-level metadata store, with individual libraries linking holdings to truly shared metadata records. (Catherine at one point said “MARC… er… descriptive records” That is, an actual slip of the tongue, not one made intentionally for effect. I suspect she intended to avoid saying “MARC” at all to avoid bringing up the issue without specifying MARC…. hm.). The example used was that 4200 libraries use essentially the same record for Freakonomics, and they all manage it seperately, and when they enhance it, their enhancements are seldom shared–and this makes no sense.
We all know this vision makes sense. We also all know that this is Ex Libris challenging OCLC’s baliwick. And, I will say, with some sadness, the fact that this vision sounds so enticing and is so different from what we’ve got is kind of an indictment of OCLC’s stewardship of what they do. This is how OCLC should be working, we all know. So Ex Libris has a vision to make it work.
How is this going to interfere with OCLC? Where are they going to get all these records from? In the flowchart of where these records would come from, libraries were identified. Can libraries legally share these records with an Ex Libris central metadata store, will OCLC let them (not if they can help it!). The screenshots of staff workflow imply that when a new acquisition comes in (or really, even at the suggestion/selection level), a match will be immediately be made to a metadata record in the central system—this implies the central system will have metadata records for virtually everything (ie, like Worldcat). Can they pull this off?
If they DO pull it off, it’ll be a great system—and will divide library metadata into several worlds, some libraries using a central metadata system provided by a for-profit vendor, others using the OCLC coooperative, others using… what? It’ll be sad to me to fragment the shared cataloging corpus like this, divide it along library ‘class’ lines, and surrendure what had been owned by the collective library community through a non-profit cooperative to a for-profit vendor.
On the other hand, lest we forget, the shared metadata world really already is divided on “class” lines–many libraries can not afford to live in the OCLC world (although I suspect those same libraries will not be able to afford to live in the Ex Libris shared metadata world either). And if OCLC actually acted more like the non-profit cooperative representing the collective interests of the library sector, as it claims to be… it would be even more distressing that it is not succeeding in supplying the kind of shared metadata environment that a for-profit vendor is envisoning challenging them with.
Oren suggested that their vision was open-ness, integration, and modularity. The implication to me is that I should be able to mix-and-match components of this envisioned URM architecture with other non-ExLibris (proprietary or open source) components.
Is this really going to be? As I in fact said last year after the introduction of the URM strategy, this URM strategy is technically ambitious even without this kind of true mix-and-match modularity. To add that in in a realistic way makes it even more ambitious. And to what extent does it really meet Ex Libris business interests (not suggesting they are misleading us about the goal, but when your plan is too ambitious to meet the timelines you need to stay in business… what’s going to be the first thing to drop?).
For instance, to go back to our Metadata discussion, if Worldcat (or LibLime, or anyone else) did provide a central metadata repository with the kind of functionality envisioned there, and further provided a full set of web services (both read and write) for this functionality… could I use Worldcat in this URM vision instead of the Ex Libris centralized metadata repository? (by 2010?). Really?
For another example, Primo is in some ways written quite satisfactorily to be inter-operable with third party products. But what if I want to buy just the “eShelf” function of Primo (because it’s pretty good), and use someone elses discovery layer with this? What if I want to buy Primo without the eShelf and use some other product for personal collection/saved record/eShelf functionality? Right now I can’t. How truly “modular mix-and-match” something is depends on where you draw the lines between modules, doesn’t it?
[If Ex Libris really were interested in prioritizing this kind of mix-and-match modularity, I’d encourage them to explore using the Evergreen OpenSRF architecture in an Evergreen-compatible way. But, yes, to do this in a real way would take additional development resources in an already ambitious plan.]
[And in Primo’s defense, if I wanted to use Primo with a third-party “eShelf”, or use the Primo eShelf with a third party discovery layer, Primo’s configuration and customizability would _probably_ support this. The former might even make financial sense with no discount—buying Primo just for the eShelf obviously would not. As more and more complex features are added, however, will they be consistently modularized to even allow this technically? It’s definitely not a trivial technological project.]
May you live in interesting times
If nothing else, it’s disturbingly novel to be presented with a vendor who seems to really get it. They are talking the talk (and mocking the interface mock-ups) that match where library software really does need to go.
Will they pull it off? If they do, will it really be open in the modular mix-and-match way we all know we need to avoid vendor lock-in and continue to innovate? Will we be able to afford it? We will see.
I would think the strength of this vision would light a fire under their competitors (including OCLC, and maybe the open source supporters too), spurring more rational vision-making from other library industries, making it more clear (if it wasn’t already) that keeping on going the way you can keep on going is not an option. I hope so. (On the other hand, Ex Libris is clearly targetting a library customer field that can afford it, in several definitionsof ‘afford’. Do other vendors think they can keep on keeping on to target different markets?]