STAFF RESOURCES

Minutes of February 26, 2015 Meeting

Present:
Natalie Borisovets, Robert Cupryk, Francesca Giannetti (recorder), Melissa Just, Triveni Kuchi, Janice Pilch, Tibor Purger, Christopher Sterback, Jane Sloan, Gracemary Smulewitz, Tao Yang (chair)
Excused:
Stephanie Bartz, Janet Croft, Mary Beth Weber
Guests:
Mei Ling Lo, Rhonda Marker, Glenn Sandberg, Jeff Teichmann, Jeremy York (Assistant Director of HathiTrust), Carla Zimmerman

Yang distributed a draft of the investigative report from the Task Force and a list of issues to be discussed with York. He suggested looking at each issue in terms of options and best practices and led the discussion that followed.

1. Shibboleth library walk-in option

HathiTrust accepts the "library-walk-in" Shibboleth attribute from partner institutions to provide certain member privileges to guest users who do not have an institutional login. Yang asked York what he knew about the practice of other institutions relative to library walk-in, and if he had a suggestion of a model to follow. York explained the set of possible services that fall under library walk-in:

Guest user access must be through Shibboleth; HT does not maintain IP addresses of member institutions. This functionality does require institutions to use the InCommon definition of "library walk-in" which is to say patron must be inside the library. HT requires certification that the access is occurring in the library. All HT needs is entitlement value that recognizes patron is in the library. On RUL end, IP address could be used to determine guest users included in this group. It is up to RUL to determine how to establish members of this group.

There are two ways to implement the "library walk-in option". RUL could create guest credentials with which public users log into Shibboleth. The other means would be determined by IP address, which would supply credentials to HT via Shibboleth. Purger said the second approach is much more workable for RUL. The drawback of this approach is that if no credentials are passed to Shibboleth, then the user won't be able to build collections (connection won't be personalized). Pilch mentioned that Dana Library is currently working out guest access to HT for a specific course. Purger said he would take charge of the IP authentication route for library walk-in users.

For visiting scholars and professors emeriti, their NetID (if they have one) will not work to access HT remotely. Yang asked if it would be preferable to provide guest login credentials. When Shibboleth access gets outside of "members", HT needs to know that there is some mechanism for addressing problems and compliance with access policies. Alumni access has been worked out at some institutions. For example, institution needs to know/intervene if user is making images from Google digitized public domain volumes publicly available. Purger suggested that one way to handle the visiting scholar/emeritus professor group is to add member value to these categories of users; or come up with another value that approaches "guest." York said access would be easy to enable; he awaits the test case.

2. Interlibrary loan and reserves

It was broadly agreed that it is inadvisable to lend HT volumes to fulfill external ILL requests. HT licenses the Google digitized volumes under specific terms to which member institutions agree. HT discourages creation of services around unlicensed uses.

Yang asked if there were differences in access between US and oversees. If professor puts HT volumes on reserve, and student is oversees, can student access those volumes? York answered that current implementation is that members should be able to access HT volumes from anywhere. There is only a difference in what US and Europe can access when member is *not* logged in.

3. Service for users with print disabilities

Yang asked if the library or the campus office of disability services should be the designated proxy in order to implement the service to users with print disabilities. York said that an institution can actually designate more than one proxies, so both the library and office of disability services can play that role, but the particular form must be signed by somebody with signing authority (Provost, etc.). Such a user must authenticate via Shibboleth. Partner institution must have policy by which print disability can be certified. Pilch raised discussion of libraries as authorized entities to provide copies to people with print disabilities (in other words, it is important for the library to have this role). Just suggested that the University Librarian should sign the form because she has a system-wide position while the office of disability services is limited to each campus.

Pilch asked what kinds of accessible copies HT provides (e.g. large print, braille). York said there were two options: [1] PDF with text layer that can be extracted (best), [2] ePub version which is text wrapped in html. So far, York has not been notified that copy received from HT is unreadable. It falls to the library or individual to extract and enlarge text.

Pilch asked if HT requires disability eligibility documents from individual users. York said no; the word of partners suffices. This falls under framework of partnership agreement. Partners assume liability for execution.

4. Discovery (API, EDS, catalog, etc.)

Smulewitz reiterated that when RUL links bibliographic records to HT, there are two options: [1] custom links for open access (OA) material, [2] OA, plus what is categorized as "restricted." Smulewitz asked if HT determines what goes under "restricted" designation. York replied that HT provides access to full text index to EBSCO, Ex Libris, and other discovery services, which essentially reuse HT material designations. Members can opt to make discoverable materials that are fully available and/or materials that are not fully available. Within discovery, if HT has determined material is in copyright, or of uncertain status, material is not viewable. Smulewitz mentioned that Fordham University only uses OA file, because everything else would require user to go to ILL and the institution wants to limit ILL requests. Conversely, MIT links to everything.

Purger mentioned that at present some materials are coming up as full-text viewable, but click-throughs demonstrate that they are not in fact viewable. Smulewitz added that the material EBSCO called "restricted" is not fully accurate either. York replied that if EBSCO says something is fully viewable, and logging into HT proves otherwise, then EBSCO is not making adequate use of HT full text index. He added that EBSCO has access to a complete file with all HT designations.

The group asked how often the full text index is available to EBSCO. York answered that EBSCO can access it whenever they wish. Notification happens when the schema changes, but otherwise it is available when EBSCO wants it. York noted that it was certainly labor intensive to perform an update for a discovery layer, but that it is entirely up to them to determine schedule by which they do it.

Sterback noted that at present the test version of VuFind only loads volumes that RUL owns in discovery layer. A preview snippet is shown, if permissible. York added that it is possible to embed HT books through discovery interface. It is possible to force login for users of discovery layer.

5. Government documents; depositing/digitizing RUL items into HathiTrust

Bartz, via Yang, asked if RUL will be asked to contribute scanned documents. York answered that there is no obligation for member organizations beyond paying annual member fee. There is no additional fee for deposit of volumes.

How does the member fee work and how does depositing factor into the fee? York replied that midway through year, HT asks partners how much content they think they will deposit. HT uses those estimates to determine how much digital storage they need to purchase. This, in association with staff costs, helps to generate annual budget, from which per-volume cost is determined. Right now, this is USD 0.16 per volume. For materials in-copyright in HT, assumption is made that value exists only where member institution holds a print copy. Partner institution only pays a portion of that per volume cost if it owns a print copy. If depositing PD volumes, cost will be lower than if contributing in-copyright volumes. There is just one member annual fee, and it is not directly influenced by the contribution of scanned volumes. Fee depends on how large member institution collection is, and on how much overlap there is with HT collection. Larger overlap usually means a higher fee.

Yang asked if HT wants to understand if contributed documents are government documents, and if state documents are part of HT scope. York answered that, yes, there is a targeted initiative to gather federal government documents. There is another initiative to determine OA status of state documents. (Note: NJ state documents do not fall into public domain by default.)

Pilch raised assumption that archival, unpublished works are not desired by HT. York replied that the general language HT uses is "book or book-like materials", which is to say bound materials that are described by MARC cataloging (not EAD). HT would say no to individual letters, but yes to compilations of letters.

York noted Bartz's question about the HathiTrust Government Documents Registry: will metadata be publicly available? York confirmed that this is the idea at this point. The Registry is a separate resource from the HT catalog.

6. Preservation

Group asked how to set up Section 108 access. York responded that access within library premises is possible (see library walk-in). HT requires information about physical condition (brittle, etc.), and/or location status (lost/missing). They then compare this to their database of out of print works. No institution has yet gone through the process of requesting a preservation copy under Section 108. York can send guidelines if RUL wishes to serve as test case. HT would open up API allowing RUL to retrieve digital copy of book and the condition would be in-library use. Is it better to print the digital copy and interfile in the stacks than to access the digital copy? Print could circulate, whereas digital copy could only be accessed in library. But the print copy would probably negate the access to digital copy because HT counts print copies.

The group asked whether there is a benefit to withdraw the brittle or damaged item from circulation, but keeping the physical artifact. Perhaps: as long as item is withdrawn from circulation, RUL should be in compliance if preservation copy sought from HT.

7. Other questions

General and specific troubleshooting: York is happy to take questions from us so RUL can take full advantage of HathiTrust Digital Library. Either email feedback@issues.hathitrust.org or York directly.

PDF text layers: The PDFs available for download from HT should have text layers (not just images). If not, contact HT at feedback@issues.hathitrust.org.

Integration with citation managers: HT currently exposes COINS metadata in page turner. Zotero can read this. Lo noted that RefWorks does not efficiently pick up citations from HT. York explained that HT hasn't supported RefWorks yet, because it relies on user login/institutional affiliation before metadata can be transmitted.



 
URL: http://www.libraries.rutgers.edu/rul/staff/groups/hathi_trust/minutes/hathi_trust_15_02_24.shtml
Website Feedback  |  Privacy Policy

© Copyright 1997-2017, Rutgers University Libraries