Saturday, November 30, 2013

And another fail scenario for EZproxy: GeoScienceWorld and the Google Maps API

This issue resolved by GSW December 8

To borrow a line from Jonathan Rochkind's Bibliographic Wilderness 'EZProxy is terrible, it's just the best we've got.' 

I don't think it's terrible but it's always been a pain that TOC services don't work off campus and that RSS feeds occasionally don't work at all, but this week I got a something new.

GeoScienceWorld's search interface includes a world map that uses the Google Maps API. -  and if we access the search page via ezproxy (and we default all users through EZProxy - even on campus) Google Maps refuses to serve data because the API Key is not registered for use in our domain (i.e. it works if you strip out the EZProxy suffix 'elibrary.jcu.edu.au')
This is the message
Google has disabled use of the Maps API for this application. This site is not authorized to use the Google Maps client ID provided. If you are the owner of this application, you can learn more about registering URLs here: https://developers.google.com/maps/documentation/business/guide#URLs
This happens as soon as you try to access GSW's search because the map is built into the search form - even if you don't want to use it.  That message appears in a dialog box - it scares off most users, who don't realise clicking 'OK' will get you to vanilla searching (sans Map) but that gives the impression search might be broken because the map disappears leaving a worryingly blank blue square.

If we remove the proxying from the link in our A-Z of database pages it will work for on campus users but off campus users won't get access to the website at all because of the IP restriction.

I haven't worked with Google Maps API - but my reading of their authorisation doco suggests there isn't a simple workaround. The prevention of other sites using your app is intentional - in a library context it's a barrier. I'm going to guess that GSW isn't interested in registering every subscriber's EZProxy domain even if it's possible.

Will be getting our people to talk to their people to see what can be done, and if anyone else has mentioned it.

People have suggested that in the future the EZProxy fail situations might be resolved by Shibboleth, or some other middle tier authentication broker. Right now I'm suggesting that power users use the VPN and strip the EZProxy suffix. Being a specialist resource most users should be easy to define (Earth Sciences), and therefore easy to inform of the problem and workarounds.

Will comment if GSW get back to  us.

Monday, November 25, 2013

Google Scholar, WoS and Informit

Not sure why Google Scholar has suddenly intersected with my job again after a seeming hiatus. The renewed activity seems to indicate Google is committed to developing and maintaining Scholar - (have already seem some conspiracy theory postings about this being the next step in googlizing the information universe now that the Google Book law suit has been dismissed)

Anyway at least I have a theme uniting a bunch of thoughts.

Google Scholar as Research Platform

I don't know when it happened but Google Scholar has upped the ante on its functionality and is horizontally integrating services provided to researchers.

They've introduced:
  • A web-based citation manager (My Library)
  • A quasi research portfolio (a list of your research output that appears in Scholar)
  • An alerting service (you get emails when new items match your criteria)
  • H5 metrics
While fiddling around I found that in settings you can configure Scholar to show 'Export to Endnote' links (also BibTeX, RefMan and RefWorks)

All you need for this functionality is a Google Account.

Google Scholar home page showing services


WoS (Wusses Out Speedily)

First there wast Thomson Reuter's (TR) announcement last week that they were going integrate Web of Science (WoS) with Google Scholar and simultaneously stop integrating it with the other 'discovery layers' (Summon, Primo and Ebscohost).

TR's email stating that this was all about improving user experience didn't gel with my version of reality and I suspected (wrongly, I assume) that this was an example of Google being evil - that TR had to be exclusive to Google to jump on their bandwagon. But a couple of days of whingeing tweets and blogs and listserv emails (and who knows what unpublic communication) from libraryland and TR announced that it would continue working with the other discovery layers. That the new announcement was so quick seemingly disproves there was any exclusivity agreement with Google - it seems it must have been a business decision to lessen their workload and thus increasing profit - pretty sure no discount would have come had they proceeded with discontinuing DL support.

Anyway, except for some raised blood pressure no harm done. Now we wait and see how Google Scholar uses the WoS data.  I'm pretty sure it will still require an institutional subscription for a user to see the data. I assume IP restriction will be used, so citation counts will appear much like our Link Resolver appears if you access Scholar on campus or through EZproxy.

Given that Scholar already has citation counts it may be that WoS will simply replace whatever Scholar is doing to create those counts.




Not so sure if users will able to attach themselves to an institutional subscription without EZproxy (like you currently can in your Scholar Settings).

Informit






The Summon developers announced that the Informit suite is now included in their unified index. I found a problem with full text links to Find It@JCU for previous titles/ISSNs, and I wanted to see if Scholar was building OpenURLs that worked - only to discover that Informit is a 'special case' in Scholar.  If you click on the title of a citation harvested from Informit in Scholar you are taken to full text in Informit - works great if you are on campus. Not sure what happens if you aren't.

I've submitted a request to RMIT Publishing asking that they work with Scholar to provide some sort of visual indication in the SERP that fulltext is available (it looks like a typical citation only hit - there is no link in the right column to either the link resolver or a URL)

Google Scholar Informit citation without full text indicator
Informit citation without full text indicator