Article Licensing Information Availability Service (ALIAS)
Mark Sullivan
Need
• identify the journal holdings in databases from member libraries
• create easy access to licensing information for journals in member databases
• double the fill rate for resource-sharing article requests within Information Delivery Services Project (IDS)
Benefit
• decrease turnaround time of article requests for patrons
• reduce workload by automating the processing required at the borrowing library
• reduce costs by decreasing the number of requests that go to libraries that charge fees
ALIAS is a system that was nearly ten years in the making, and it is now in use by 70 IDS Project libraries. The IDS Project was conceived by Edwin Rivenburgh, then director of Milne Library at SUNY Geneseo, and it was backed by eleven other New York State academic library directors. These directors believed that to survive and prosper in the twenty-first century, libraries would need to radically change the way they do business with each other. Towards this goal, the IDS Project was developed to advance the sharing of library resources through collaboration, innovation, and efficiency.
Members commit to meeting delivery goals based on a user-centric definition of an interlibrary loan transaction: from the time the user places a request until the time the user is notified the loan is ready for pickup or is ready to be retrieved from the Web. The IDS standards for delivery are 48 hours for articles and 72 hours for loans. To date, the IDS Project has over 70 member academic, special, and public libraries, including the New York State Library and the New York Public Library.
In the fall of 2004, soon after the IDS Project began, we ran a newly developed statistic system to determine the transaction times and fill rate within the member libraries. While the overall statistics showed that we had a lot of work to do to meet the delivery standards, the results were very promising. However, the fill rate for articles within the project was exceedingly low, and it seemed impossible that we were filling less than 33 percent of the article requests considering the large amount of money each library was spending on journals. All member libraries were using ILLiad and OCLC for resource-sharing management, which worked extremely well for monographs and print journals. The problem was that none of the libraries were loading holdings information for electronic journals into OCLC. In their defense, there wasn’t a good way to do this.
To test the theory that we could fill a much larger number of article requests with electronic journals, we cobbled together a web-based search system that ran independent searches of each library’s OpenURL resolvers, which contained their journal holdings data. The results were dramatic and showed that we could fill another 31 percent of article requests from within the project if we could discover what each member held—and if we could determine the licenses for each one. In 2005, SUNY switched from Ex Libris’s SFX to Serials Solutions, which simplified the process of extracting holdings data because all of the other non-SUNY IDS libraries were using Serials Solutions.
After several months of unsuccessfully searching for an electronic rights management system that would suit our needs and have some general licensing data already loaded, we built our own. Data was downloaded from Serials Solutions and loaded into an Index Data Zebra z39.50 index server. The staff members at the project libraries were able to connect to our newly created eJournal Availability Server using the ILLiad client and to run searches on the ISSNs of the journals. The first stage of development was never implemented on a project-wide scale, however, because no one was comfortable loaning articles without knowing that the contracts allowed it.
The second stage involved obtaining license and contract information for the interlibrary loan (ILL) rights of each electronic journal. Instead of using the traditional “What did you sign at your library?” we decided to see what each publisher and aggregator was displaying on their websites. General terms of service, user rights, and licensing data were collected from the websites of nearly all publishers and providers. Publishers that didn’t have information published on their websites were sent e-mails requesting their general contract language. Interestingly, several responded that they had not considered ILL rights for e-journals and asked if we would provide language for them to include. The language we provided was the equivalent of Fair Use Rights. It took several months to compile the licensing data and to integrate it with the holdings data. The first step of discoverability and license availability was completed, but it was done manually, which was slow.
In 2008, the IDS Project began working with Atlas Systems, the creators of ILLiad, to produce an improved version of the existing e-journal availability. This new version would have several components:
1. Similar holdings and license database
2. A load-balancing web service that would spread the requests among all of the members
3. An ILLiad server-side Microsoft Windows service that would allow for automatic request processing based on the holdings, license, and load-balancing data.
Together, these three pieces would form the Article Licensing Information Availability Service (ALIAS). The IDS Service and load-balancing system that Atlas System’s created took our data system and turned it into a fully unmediated article request system.
ALIAS was beta tested during the spring of 2008 with six libraries and brought online for all IDS libraries after the August 2008 IDS Project Conference. One of the most immediately noticeable results of the unmediated service was the decrease in time between receiving a patron’s article request and sending the request to OCLC. Prior to ALIAS, it was not uncommon for a request to sit in queue for several hours or even more than a day. During the initial beta period of ALIAS, some of the libraries were able to demonstrate the improvement by reducing the average time from more than a day to an average of approximately two hours for the entire request. The data also showed that even with the amount of effort put into ALIAS, only 54 percent of the requests submitted could be processed automatically. We found that
• 44 percent of requests were sent to OCLC and then on to the first library in the ALIAS-generated lender string;
• 10 percent were determined to be held locally and were routed to a local holdings queue; and
• the remaining 46 percent were handled manually. The IDS Service placed a note in each request explaining why ALIAS was unable to process the request.
The majority of nonlicensing issues were due to a lack of an ISSN in the request. The remaining issues either involved a license that didn’t allow for loans or no record of the journal being available at any of the IDS libraries. Considering that only six libraries were involved with the beta test, we hoped that this number would decrease substantially once the rest of the libraries were brought online.
The data from January 2013 to May 2013 has shown that 50 percent of over 60,000 article requests were handled successfully by ALIAS and another 19 percent were handled manually and sent to other IDS member libraries. The initial issue was why the 19 percent failed to be processed automatically. The primary reason was that patrons were submitting invalid ISSNs. Over 9,000 of the unsuccessful requests had invalid data, including an apology in the ISSN field for not having the proper ISSN. Better discovery tools and user education would help reduce this number substantially.
The remaining 31 percent of the requests that required going to libraries outside of the IDS Project were due to a lack of holdings. However, it wasn’t that the member libraries didn’t have the required journals; they didn’t have access to the required issues. The electronic journal embargo periods and lack of digitization of older volumes prevented the member libraries from filling over 8,000 requests (13%) during the first half of 2013. If the embargo period was removed so that e-journals were treated like their print counterparts, 82 percent of the requests would have been filled. The remaining 18 percent would still need to be filled by libraries outside of the IDS Project or through purchase elsewhere.
The ALIAS license data is updated each year by a team of 20 or more librarians from across the IDS Project libraries. However, spring 2013 marked the first time in years that the interface to the ALIAS database was updated. A joint librarian, faculty, and student team at SUNY Sullivan created a new interface that should speed up the processing of the licenses dramatically.
In the next few months, the IDS Project will begin work on ALIAS 2.0. This new version is based upon ILLiad 8.4 and the server-side Addons that were made available in version 8.3. ALIAS 2.0 will begin with the conversion of the IDS service from a Windows service into an ILLiad server-side Addon. This change will
• allow easier installs and updates,
• allow for easier searching of Serials Solutions/SFX for missing ISSNs, and
• no longer require Atlas to install the Windows service onto the OCLC host servers or campus information technology departments to install it onto local servers.
ALIAS 2.0 is currently under development.
The IDS Project is committed to providing improvements to the entire resource-sharing community through the ongoing development and sharing of innovative tools and promotion of best practices. ALIAS, although primarily a system for IDS Project Libraries, has also had an effect on outside libraries through the freely accessible Serials Solutions and SFX Addons. ALIAS also provided the model for OCLC’s Direct Request for Articles, which is now in its third year of operation.1 The IDS Project will continue to work with Atlas Systems and OCLC to improve workflow and to create new tools for the resource-sharing community.
1. “OCLC, IDS Project and Atlas Systems to Develop Network-level Solution for Unmediated Article Resource Sharing,” http://worldcat.org/arcviewer/1/OCC/2011/11/08/H1320767091756/viewer/file4022.htm.