Scot Dalton at NYU continues in his impressive tradition of adding apparently well coded features to Umlaut without talking to me about them at all while he’s coding them. I’d like to think this is because the Umlaut code is so well-written and clear, but it’s probably more that Scot is apparently a programming genius.
Scot tells us he’s added an adapter (basically what we call Umlaut ‘plugins’ for different external services) to Umlaut to display Ex Libris bX recommendations:
I just committed a bX recommender service adapter to the trunk.
Here’s an example:
Let me know if you have any questions/issues.
We don’t subscribe to bX where I am, so we won’t be using thi, at least not yet. We DO provide somewhat similar (although probably not quite as good) ‘more like this’ links to pages in Scopus and ISI. Scopus/ISI ‘more like this’ aren’t based on usage data like bX, but on metadata similarity instead. Umlaut can also provide ‘cited by’ links in Scopus and ISI, which can serve a ‘more like this’ functon.
If I were using Bx, I’d want to combine it with my own Scopus/ISI links, so I’d probably want the UI to be a bit different — having the Bx results be on a different page or an AJAX popup, so the Bx option could be presented in the same list as Scopus/ISI. But I trust I could enhance Scot’s code to do that if needed.
One issue with Bx and Umlaut, that I don’t believe Scot/Ex Libris have solved yet — SFX itself has a bug (or lack of feature), where ‘click through’ statistics are missing/corrupt/useless for SFX usage through the API. If you’re using Umlaut, all your SFX usage is through the API, so SFX’s own click-through statistics are useless. Bx can take your usage statistics and use those as part of it’s data for calculating usage-based ‘more like this’ — or optionally as ALL of it’s data, limiting ‘more like this’ calcluations to be based only on your own institution’s usage. However, this is, as far as I know, not available if you use Umlaut, or any other SFX API usage, as your main SFX usage. That’s really up to Ex Libris to fix on the SFX side, although another possible solution could be having Umlaut track it’s own usage data, if Ex Libris would accept usage data in a standard format from software other than SFX, something they’ve talked about doing but I don’t believe they make available yet.
But even if you can’t send your usage data to EL, you can still use the Bx service based on others usage data, which I’m guessing is what NYU is doing.
Kudos to Scot. And extra kudos to Scot for contributing this code back to Umlaut, this is now easily available for anyone.