SFX Documentation

ScienceDirect now sports bX recommendations

bX recommendations for scholarly articles, a.k.a. Get It Recommends, are now available in ScienceDirect for K-State users. Ex Libris announced this extension of bX services back in October; read the press release.

The recommendations appear in a box on the right side of the ScienceDirect display when the user is looking at an article's detailed view. The service is available whether or not K-State has access to the article's full text.

(click image to enlarge)

Based on the article being viewed, bX may or may not offer recommendations for other relevant articles. If no recommendations are available, the box will display the text, "The record has no recommendations." If recommendations are available, they will be listed in the box and will link to the referenced article(s). Initial testing indicates that the recommendations appearing in the box link to articles that were originally published by Elsevier.  This functionality is different from bX's Get It Recommends service in the Get It menu in that the recommendations produced by Get It Recommends are publisher-independent. For more about Get It Recommends, check out this K-State Today article.

bX recommendations inside ScienceDirect are available to all on-campus users of K-State's ScienceDirect access. Unfortunately, off-campus users do not see these recommendations right now. The service has a known issue with EZproxy that prevents it from displaying through a proxied connection. bX customers who use this service (including K-State Libraries) are lobbying Elsevier for a fix. This problem is internal to ScienceDirect and does not affect the Get It menu's display of Get It Recommends.

SFX's new embargo behavior

An e-journal embargo can be defined as the elapsed period of time between a journal article's publication and its availability through an online interface. The presence of an embargo affects whether or not an article is accessible to the end user requesting it, so SFX keeps track of embargo periods as part of its threshold information. Sometimes SFX faces a situation in which it cannot accurately calculate whether the full text of an embargoed article will be available to the end user. It is in these ambiguous situations that SFX's embargo behavior has changed.

Working the workaround: DTFT Local

Making SFX link to dissertation and thesis full text objects is a big project. It's fairly easy to get SFX to do something with the citation; it gets more difficult when that "something" involves consistently delivering the user to the full text.

Many dissertations and theses are offered as electronic full text through the product ProQuest Dissertations & Theses Full Text (DTFT). The global DTFT target requires "genre=dissertation" to be present in the OpenURL but most A&I databases don't index dissertations this way. Rather, they index the contents of Dissertation Abstracts, presumably because this is a comparatively easy way to acquire a broad swath of dissertation and thesis citations. The result of this practice, however, is that when these A&I databases are SFX sources, they transmit OpenURLs that mark up dissertation citations as if they were journal articles.* You can make full text dissertation linking work around this problem by fooling SFX into thinking that it's linking from journal article citations to journal article full text. Many SFX customers have already done so and, as of Monday, K-State is one of them.

Scripted bliss

Today I discovered the wonders of the change_availability script. Even though I sometimes feel like an old hand at SFX, this feature was new to me. It's a wonderful thing; I hope you get to know it sooner than I did.

I have neglected the new free full text targets for a while now and was facing the ugly, time-consuming task of activating them, their object portfolios, and getFullText services by hand. Locate the targets that use the getFullText service; figure out which targets have object portfolios; click, click, click endlessly through SFXAdmin to get everything activated. Ugh.

Notes prior to ELUNA

This week I'll be attending ELUNA, the Ex Libris Users of North America conference, gathering ideas and news about SFX. I hope to hear some chatter about a topic that's flared up on SFX-DISCUSS-L - that of article-level linking to ProQuest Dissertations & Theses Full Text. Last week I learned that the Help Desk staff have for some time dealt with faulty linking to this target as a "known issue." Yikes. I can manually edit the OpenURL and force the link to succeed, but that's my job. You shouldn't have to know how to do that.

Very brief investigation has shown me that some of the fault, dear Brutus, lies not in our target but our sources. CSA PsycInfo, for example, sends the thesis or dissertation citation to SFX encoded as an article (which, from PsycInfo's point of view, it is, but that's a different post). The ProQuest D&T target will only accept a request that's encoded as a dissertation. This miscommunication is a symptom of a larger problem, but take the nutshell for now.  I want to explore how the fixes propagated on SFX-DISCUSS-L work - whether they truly repair the miscommunication or just spackle the holes. I should learn more this week - then prepare for the ugly post on why all this miscommunication is happening in the first place.

A couple of unrelated notes:

• Shiny gold stars and happy smiles go out to Inga Overkamp at the Max Planck Digital Library, for sharing her temporary solution to Synergy Blackwell / Wiley Interscience linking woes. I undid her fix today with a bit of sadness because it served us so well while we needed it.
• A welcome mat goes out to Sarah Miller at CARLI, who drank the Kool-Aid and joined me and other chatty SFX admins in blog-land. Welcome!

IEEE Xplore Journals & issue-level linking

Thanks to a problem Laura reported earlier this week, I've temporarily turned off CrossRef lookups between SFX and IEEE Xplore Journals. This means that until I turn it back on, this target will only resolve to the issue level, no matter how much article-level data you give it.

To understand why on earth I'd do this, you need a little backstory.

Knovel activated

I have activated object portfolios for our subscribed ebooks in Knovel.
A total of 1520 objects were activated in this process. Knovel creates different objects for the same work published under different labels, so this number does not reflect the number of active unique titles in the database.
I will be testing this target as it rolls to production. Please let me know if you spot any abnormal behavior. I always appreciate other sets of eyes.

JSTOR Display Logic

I've implemented display logic for JSTOR so that an object available from multiple JSTOR collections will only show up once in the SFX services menu.
I started with rules for only Arts & Sciences 1-4, hoping that these would catch the majority of overlapping titles. If you still find problems with overlap, I will play with the logic rules some more.

Ulrich's target service

Jason sent a weirdness email on 1/30/07 that I replicated. He invoked the SFX services menu in Ulrich's for a journal title "The Nation" from Malawi with no ISSN. The menu gave him 1 choice, this time with an ISSN, to ProQuest. The journal title "The Nation" in ProQuest was from New York, and had an ISSN.

The Ulrich's target service was at fault here - ULRICH::ULRICH
The global threshold was  $obj->need('ISSN') - requiring an ISSN to hook to the KB holdings.
That explains why it excluded valid titles with no ISSN, but not why it only grabbed one "The Nation" with an ISSN. The New York title is the first in a list of 2 or 3 "The Nation" with ISSNs - why not invoke the multi-object menu?

Anyway. I added a local threshold that accepts a journal title: $obj->need('rft.jtitle')
and allowed SFX to use either global or local threshold in the Ulrich's Target Service:
GLOBAL || $obj->need('rft.jtitle')

The problem should be solved, but I still don't know why SFX didn't invoke the multi-object menu...

'Public Name' attribute

After editing both the Target and Target_Services windows for MIT Press, I saw that the "Get It @ K-State" window displayed it as:
MIT Press: MIT Press
instead of the usual:
Get full text at: MIT Press
Turns out (thanks, Dale), editing the 'Public Name' attribute in the Target window is mandatory, and results in the bold service name after the colon.
Editing the 'Public Name' attribute in the Target_Services window is not mandatory, and results in that text masking the name of whatever service is being offered (ft, abstract, etc). Seems obvious now, but wasn't at the time.

MIT Press is now fixed, but there is a larger use to which this masking can be put. Should we want to address the users in a more colloquial fashion (again, thanks, Dale), we may want to add public text to the Target_Services window. The best example of the moment is LexisNexis:
Continue your search at: LexisNexis
(or whatever text is deemed appropriate), since users can't even link at a journal level in this DB.

The other issue is the near-future task of shortening the public names of many of the DBs in the Target window. A user will see
Get full text at: JSTOR
rather than the current version:
Get full text at: JSTOR ARTS & SCIENCES 1
I will be asking Carol about queries to extract these titles when I go to Boston next week. Stay tuned.