[Web4lib] FW: Troubleshooting Google Scholar Resolver, JSTOR metada and 360 Link

[Bob Duncan]

At 07:49 PM 4/22/2009, Dyer, Renata wrote:
>We subscribe to Serials Solutions 360 Link service and have, amongst 
>other resources, linked JSTOR  and Google Scholar holdings to it.
>
>The problem we are experiencing as of February this year is that 
>when JSTOR holdings are recognised by the Scholar, the 360 Link is 
>bringing up a citation screen containing only title and a weird 
>looking DOI and and error message saying DOI is invalid. JSTOR 
>support confirmed that the Scholar is misinterpreting article 
>metadata (read: DOI) and that they will try to resolve it with 
>Serials Solutions and Google.
>
>In the meantime my patrons are clicking on the Google Scholar link 
>that shows an error citation screen and does not take them to a full 
>text as it should! Of course if they click on the linked title 
>they'll go straight to the JSTOR article but we tried so hard to 
>train our client to use the Findit link in Scholar every time they 
>see it  and now they are having this very frustrating experience.
>
>Is anyone else experiencing problem with JSTOR holdings and other 
>Open URL resolvers?


Short answer:  no.

Long answer:  I'm not sure I understand the problem, and the above is 
confusing on many levels, but it sounds like your problem has little 
to do with JSTOR and mostly to do with your 360 Link 
implementation.  (Assuming 360 Link works like most other OpenURL 
resolvers and SS hasn't cut some special deal with GS.)

-  Google Scholar does not "recognize" JSTOR or any other holdings 
you may have.  The institutional_holdings files you provide to Google 
Scholar indicate journals and dates, but not providers, and is used 
by GS soley in the determination of whether or not an OpenURL link 
will be generated for your users.  GS has no knowledge of your JSTOR access.

-  Clicking your Findit link sends the OpenURL generated by GS to 
your implementation of SS 360 Link, so what happens in the 360 Link 
window is a result of what 360 Link does with the metadata in the GS 
OpenURL.  If you've told GS you can deal with DOIs, the OpenURLs you 
get from GS for articles that have DOIs tend to have the DOI and 
little else.  It's up to your OpenURL resolver to deal with this 
situation; either it needs to be able to link to the article using 
the DOI provided, or it needs to submit the DOI to CrossRef (or 
whatever), get metadata back, and use that metadata to create links 
to the article.

-  In my experience, GS does not include "weird looking" DOIs in its 
OpenURLs, but it does create version 0.1 OpenURLs, so if 360 Link is 
expecting to see (e.g.) rft_id=info:doi/10.1063/1.44091, but it gets 
id=doi:10.1063/1.44091 instead and doesn't know what to do with it, 
you may indeed get something odd looking in place of a valid DOI.


Can you provide an example search result in GS where the DOI seems 
off and where you see odd behavior that you think is JSTOR-specific 
and not related to Google Scholar's horrible implementation of 
OpenURL or 360 Link's possible shortcomings?  (And if you could 
provide the Google Scholar OpenURL for the result too, it would be 
helpful to see what's going on with your resolver.  The library link 
for "Treasury Department (Australia)" in GS preferences is not 
selectable from here so it's impossible to see what you're seeing.)

Bob Duncan



~!~!~!~!~!~!~!~!~!~!~!~!~
Robert E. Duncan
Systems Librarian
Editor of IT Communications
Lafayette College
Easton, PA  18042
duncanr at lafayette.edu
http://library.lafayette.edu/