Home » OneSearch (Page 5)

Category Archives: OneSearch

OneSearch vs Aleph as an E-resource Source

2015-11-11 / Leave a comment

Last week’s post about link checking describes one of the ways that OLS is working on improving access to e-resources via OneSearch.

Another area of work regards e-resource records batchloaded to Aleph which are also available via the Primo Central Index (PCI). Because of the large number of duplicate records coming from Aleph, there are often problems determining availability correctly for these batchloaded records.

To avoid these problem, when OLS finds that PCI records are duplicated in Aleph, we mark the Aleph records in order to prevent transfer of these records to OneSearch.

Note: OLS is not removing records from Aleph, we are simply not using those records in OneSearch.

Additional benefits of this approach:

Efficiency. Duplication doesn’t only happen with PCI. Sometimes the same records are loaded to Aleph for multiple libraries. Any library using OneSearch as their discovery tool can choose to stop running these duplicate batchloads.
Faster turnaround with updates coming to PCI directly from the publisher.
Reduced potential for error and out-dated content.

How do I know which Aleph records are being blocked?

STA = PCI means we don’t need this record in OneSearch. These records can be found in Aleph via a CCL search (where “xx” is your library code):

wst=pci and wow=xx

What if there is more metadata in the Aleph record than in the PCI record?

Many PCI records come with searchable full-text — for example, in addition to the metadata, the entire text of an article could be searchable. Although it is possible that PCI records will contain less metadata than Aleph records, in such cases, we believe that the searchable full-text is a more than ample substitute. On the OLS Support Site, the following information is available:

How do we know if specific e-resources are available in PCI in order to decide what content is duplicated?

PCI makes activated collections searchable (aka “discoverable”). Full-text access for CUNY (aka “Full text Available” in OneSearch) happens because the resource is also activated in SFX. Many PCI collections can be searched for free, even when delivery of full-text is only available to subscribers. See the Primo, PCI, and SFX blog post for more information on this relationship.

At CUNY, we have activated all free-to-search PCI collections in our testing (“Sandbox”) version of OneSearch. When we want to check if specific e-resources are available via PCI, we can search for them in the Sandbox and, after the initial search result is displayed, we click on the “Include results without full text online” checkbox to include content to which we do not have full-text access.

Aleph e-resource records that are not found in PCI will continue to be sent to OneSearch.

Sandbox Caveat: The Sandbox does not include all CUNY content. It is restricted to 100K (local) records so it only contains a small percentage of our Aleph content.

Why does OneSearch sometimes say “No full-text” when there is a working link to full-text in the “Details” tab?

Distinguishing what is really available as full-text is challenging. To minimize user frustration, it is considered preferable to under-promise and over-deliver rather than to over-promise and under-deliver. In other words, the system errs on the side of false negatives (“No full-text” displayed even though full text is available), instead of false positives (“Full text available” displayed even though full text is not available).

For this reason, it can be helpful to remember that while “Full text available” is generally reliable, “No full-text” is less reliable. It is usually worthwhile to click on the “Details” tab of a “No full-text” result, just in case it contains some link that will get you to full-text.

Example: HathiTrust provides no indication of whether its content is open access or not, forcing its users to choose between false negatives and false positives. In response, we at CUNY have chosen to indicate that all HathiTrust content is “No full-text” even though many records will contain a link to full-text in the “Details” tab.

Are e-books showing up in Google Scholar? Many faculty will go to Google Scholar as their discovery tool and bypass the library.

E-resources activated in SFX show as available at CUNY in Google Scholar. This means that a side effect of this project is more availability information being shared with Google Scholar, inadvertently leading to a better user experience in Google Scholar. Please be aware, however, that Google Scholar shares little about how and where it gets the information it provides.

What are e-resources?

“E-resources” refers to any type of content that is available in electronic or digital form. This includes articles, books, videos, and more.

New ways to search OneSearch

2015-09-14 / Leave a comment

Over the Labor Day weekend, the Office of Library Services added a handful of new fields to the search section of the PNX for records retrieved from the CUNY Catalog: MARC 5XX and MARC 250. (This was determined based on a survey OLS distributed to the Cataloging Committee and Public Services Committee in July 2015.) In layman’s terms? Your users will find more local results in their searches!

In technical terms, we’ve made the following MARC fields searchable in OneSearch:

250: Edition Statement
500: General Note
501: With Note
502: Dissertation Note
508: Creation/Production Credits Note
511: Participant or Performer Note
522: Geographic Coverage Note
546: Language Note
550: Issuing Body Note
585: Exhibition Note
586: Awards Note
590: Local Note

Users will now find videos when searching for an actor, specific editions of textbooks, items in a special collection, and more! To facilitate this, we’ve also included many more of these 5XX notes in the “Details” tab of each record with appropriate labels:

500 (no special label, other than general “Notes” field label)
501: With
502: Dissertation
504: Bibliography
506: Restrictions
507: Scale
508: Credits
511: Performers
515: Numbering
516: File Format
521: Intended Audience
522: Geographic Coverage
525: Supplements
530: Other Formats
533: Reproduction
534: Original Version
536: Funding Source
538: System Requirements
540: Use Restrictions
545: History Notes
546: Language
547: Former Title
550: Issuing Body
555: Finding Aids
580: Related Works
581: Publications About This Work
585: Exhibitions
586: Awards
590: Local Note

In the following example, you can see one of the results for the query “Fraud and Swindles Collection” in the John Jay instance of OneSearch:

Note that the “Local Note” field is not only searchable (as it came up for our exact phrase search) but is also displayed. In this way, users will know that they can search for other records in a similar way.

We hope these enrichments to the display and searchability of CUNY Catalog records in OneSearch will greatly improve the experiences of your students, staff, and faculty! Please let us know if you have any questions or concerns.

Personalize your OneSearch results!

2015-09-11 / 5 Comments on Personalize your OneSearch results!

As of mid-August, OneSearch users have the ability to personalize their results right from within the search screen. (Previously, this was only possible from within “My Account” in the user’s personal settings.) What does this mean? Basically, users can provide context for their queries.

The ambiguous search term “magnet,” for example, provides over 66,000 results:

Without any personalization, OneSearch doesn’t know which results are more relevant so it floats reference entries to the top of the result list. However, with some context, OneSearch is able to recalculate relevancy and re-position the results so the more relevant results appear on the first page of results.

For example, that same search for “magnet” within the context of the “Sciences” floats certain results to the top:

Within the context of “Education,” though, other results for “magnet” (previously buried within the thousands of other results) appear at the top:

In addition to choosing a discipline for personalization, there is also an option to float newer results to the top for cases where that makes sense:

Screenshot of CUNY OneSearch's 'Personalize Your Results' pop-up menu

The results themselves never change, assuming the query remains the same. Personalizing results, instead, changes the relevancy of the returned results. This applies, therefore, only to results that are sorted by relevancy—which is the default sort order in OneSearch.

Another nice bonus of this updated OneSearch feature is that it clearly states, at the top of a user’s results, that the records the user is seeing are personalized. If the user wants to change the context or remove the personalization, it can easily be done from the results page:

Screenshot of CUNY OneSearch 'Personalize Your Results' links on the results page

We hope this helps your users find more relevant results in OneSearch!

Primo: Palgrave Connect

2015-09-08 / 1 Comment on Primo: Palgrave Connect

The CUNY-wide licensed ebook titles of the Palgrave Connect collection are now active in SFX and can be discovered in Primo.

In SFX, 6,491 portfolios are active for all of CUNY in the PALGRAVE_CONNECT_EBOOKS_COMPLETE target.

Impact of Primo on SFX

2015-08-27 / 2 Comments on Impact of Primo on SFX

A few weeks ago, we (OLS staff) discussed ways in which the impact of Primo on other systems and environments could be illustrated.

A good place to start with is SFX, since it’s so heavily involved in creating fulltext links in Primo. So one way to measure Primo’s impact on SFX is to count the number of OpenURLs generated by Primo (either through “View Online”, or through the menu under “Other Options”) and compare them to other providers/generators of OpenURLs.

In FY 14/15, the top 5 providers were Primo, bX, Ebsco, Google, and Gale (For the remaining providers, the numbers drop significantly). Here’s a table with the number of OpenURLs generated each month by each provider:

                bX	Ebsco	Gale	Google	Primo	Other
July-14	        3052	14185	6990	6847	308	9672
August-14	1147	7512	1571	5821	859	12186
September-14	4477	25903	3289	10940	19302	25912
October-14	10651	44538	8009	16150	63698	35567
November-14	12934	49530	11050	17189	97932	25718
December-14	9506	31664	7530	15206	86557	19483
January-15	2583	9595	2030	6925	22495	12034
February-15	6194	24143	3317	9526	49877	17518
March-15	12127	41397	7257	15032	96998	24205
April-15	11756	40696	8220	14330	103400	23534
May-15	        10033	33960	7951	13107	99598	18721
June-15	        3494	14075	1955	7026	28637	11702
TOTAL OpenURLs: 1538333

I thought the best way to show the behavior over time is to use an area chart, because it shows how values (# of requests) change over time for different categories (OpenURL “generators”):

Since it’s implementation at the end of August 2014, Primo has quickly taken over a large share of the OpenURLs resolved by SFX, at a small expense to the other providers, but not by much. It actually looks as if the percentage of requests for the other main providers has remained relatively stable

The same chart in absolute numbers looks like this:

So it looks as if Primo has taken over a large share of the OpenURL requests, but not so much by taking them “away” from other providers but by simply increasing the number of OpenURLs.

The total number of SFX requests for FY13/14 was about 1.1 million (1,107,169). In the past FY, this increased by almost 40% to 1.5 million (1,538,333). And of these, 670K (669661) came from Primo.

Observations:

In FY 14/15, Primo was responsible for over 40% of OpenURL requests, i.e. SFX menus
It reached this number very quickly, within about 2 months of going live
It simply increased the number of requests, and did not diminish the contribution of other vendors
Given the impact on SFX, the impact on e-resource usage (search sessions, etc.) should be significant as well

What's New @ OLS Systems