So Ebsco and Ex Libris used to be friends. Ebsco used to allow their content to appear in Ex Libris’ pretty awesome index of academic stuff (aka Primo Central a webscale Discovery service). Then Ebsco decided they wanted to create an equally awesome index and took their data out of there, as well as out of everyone elses. Others probably did the same in retaliation. Libraries as customers of all parties are unhappy about this, and a major US consortium, the Orbis Cascade, has asked both to shape up.
I’ve spent a couple of years looking at webscale services. There is a lot to like about all the major products on the market, so I’ve no personal favourite in any of this. I think that in general they are a good thing for libraries, if not the magic bullet that some vendors make them out to be. Their main use case is probably for access over discovery by provding links and classmarks for a libraries’ inventory, but that is still pretty useful.
Regardless, its a shame about this whole data sharing mess. Adrian Phol over on the OKFN Open Biblio blog has left a concise summary of the tricky situation, and a suggestion that getting open data out there the right way to go. In principle, I think he is right. But for now, metadata = money for all of these guys (even Ex Libris), and the financial advantages of free dissemination of descriptive data have yet to be realised in the academic publishing board room. I’m not sure that libraries can afford to wait until this attitude changes.
What can we do as customers?
Tell the vendors to sort it out. But talking to your friendly product rep may not help, they have little swing in these situations. The time to handle it is at contract negotiation, we need to buy back the right to to make the content we’ve already licensed discoverable via whatever tool or mechanism we choose.
In an ideal world, we would be with-holding 5-10% of our subscription payments for licensed content until useful data is in place in our preferred discovery service . There is a sound reasoning behind this. A subscription is only partly useful if we cannot make it work within and as part of our own web presence. That means our discovery services and link resolvers. We could also throw in coverage in a ERM (with license details), useful deep linking, working federated access and a working proxy config as well.
Why should we pay full price for an incomplete product?
Realistically, this will only work if we scale it up and request en-mass. Purchasing consortia such as Jisc collections can play a role here.
What can vendors do?
Sort it out. If you cannot do it bi-laterally, ask one of the trade bodies to step in, the Publishers Licensing Society might be a good fit. Let them set up a neutral clearing house of data. you don’t need to give full text, or even abstracts, although it would be better for us and you, as it will ensure more of our users get to your content, and that means we are more likely to renew our subs for your stuff.
Are API’s really better than pre-indexed content?
This is curious. Ebsco have made claims along these lines to justify the API only access to their service, and Ex Libris still support this model of access through their Metalib+ product. Relying on APIs to replace pre-indexed content works to a point, but does not scale. Merging 1-2 data sources is fine. 10-12 gets silly. And we’ve tried it before. Its called Metalib, or Webfeat, or 360 Search, or Sirsi Rooms. And we need to step away from that, like the rest of the web did 10+ years ago.
EBSCO Discovery is itself a large index (and a very well crafted and presented one at that). So their arguments over the advantages of API’s seem mystifying.