No one said it would be easy

@laytor has run into early problems trying to implement the API.

Having followed the documentation supplied by Thomson Reuters for the API it is throwing an error that we have been unable to interpret due to lack of specific expertise in Java. It seemed, in fact, that the documentation we have might not be up to date as the terminology seems to be a little at variance with that used in the recent webinar (Introduction to Thomson Reuters Research Evaluation Tools) that @laytor and I attended.

I’ve been in touch with our contact at Thomas Reuters who have confirmed that the documentation is indeed a little out of date and they are on the verge of releasing new versions of many of their WS documents, hopefully by the end of the month. To be fair, I did harangue them somewhat to ensure we had enough info for #jiscri bid and I’m grateful they sent us what was available so quickly.

We are also fortunate that @laizydaizy agreed to help us out with her Java skills even though she is not on the project tem and very busy with the JISC funded, repository-related PC3 project – we have worked closely together before of course on earlier JISC projects – an obvious benefit of networks that have been forged by these projects (though we’re also institutional colleagues of course!)

We are also having a little difficulty disambiguating the various services offered by TR and how the API fits in. The webinar referred to WoK web services lite and premium as well as Article Match and Retrieve and researcher ID.

Here is a summary to my naive questions from the nice man at TR:

Article Match Retrieval service has one very specific purpose: it takes in metadata and pitches out URLs that can be used to link to specific data elements in Web of Science and Journal Citation Reports. As far as I know, you haven’t been entitled to this service, but it would be easy enough to arrange. The service is free.

Web Services Lite: This service responds to queries to return a limited range of data elements from the Web of Science. The fields are Author, Source (volume, number, issue, date, page span), Article Title, Keywords, and UT (a unique record identifier). The primary use case for the Lite service is to populate institutional repositories and is scheduled to be made available within the next two-to-three weeks. This service is also free.

Web Services Premium: This service is a much more robust version of WS Lite and is very similar to the API we sent you earlier. The primary differences are that the service needs to be entitled and has much, much better documentation. WS Premium is scheduled to be available within the next month to six weeks. I’m not sure what the price is (if any) for the service, but we hope to have that sorted out in the very near future.

Though I was present while @laizydaizy and @laytor talked about Java classes and suchlike I didn’t understand a whole lot (to be honest it was like being at some sort of Kabbalistic cult meeting) but I think the gist was that there may be a fundamental problem with the code generated by TRs batch file itself – @laizydaizy got all excited as she pored over the esoterica on @laytor’s double screens (I want double screens!)and put it all on a usb drive to take back to her lair. She’ll let us know how she gets on when she emerges.

The other issue she raised was that the API appears to be using an older version of Java and it remains to be seen to what extent the new documetation is updated. This might mean that even if we do get the thing working we’ll end up with what is in effect legacy code – this is something else we need to clarify with TR and feed back to JISC.