Discovery Interfaces: The Next Generation<br />To boldly go where nOPAC  has gone before…<br />Amy Donahue<br />Project Le...
Outline<br />Objective<br />Motivations & Goals<br />Introduction<br />Methodology<br />Literature<br />Interview Results<...
Objective<br />To perform an environmental scan of next generation discovery interfaces. <br />
Motivations & Goals<br />LO Strategic Plan on Collection, Preservation, and Access to Information<br />Update older system...
An Introduction to Catalog Interfaces<br />Front end vs. Back end<br />Public interface / ”librarian” modules<br />“We sho...
Methodology<br /><ul><li>Literature
Interviews
Web </li></li></ul><li>Literature<br />Finding:<br />Library database search: <br />Ovid – Wilson Library Literature<br />...
Interviews<br />Selection of candidates (in order of interview)<br />NLM: Wanda Whitney, Dee Clarkin, John Doyle, Dianne M...
Websites<br />Finding:<br />Word of mouth<br />Conferences<br />“Berry-picking” (linking to different sites from original)...
Results<br />
Results<br />Collection of 105 resources<br />Bibliography of 16 resources<br />Recommended reading list of 99 items<br />...
Discussion &Recommendations<br />
Discussion points<br />The most pervasive theme found during the scan was the call to design library interfaces that are i...
Recommended Features<br />Advanced search<br />Facets<br />Visuals<br />“Did you mean”/spell check<br />Inclusion of multi...
Recommended Interfaces<br />
Other Interface Considerations<br />Accessibility (508)<br />Mobile interfaces	<br />Language<br />Content visibility<br /...
Other Examples<br />World Digital Library<br />National Library of Australia Prototype<br />Dartmouth: Summon<br />America...
Recommended Actions<br />Evaluate<br />Interface options (cost-benefit analysis)<br />Users<br />Chosen interface<br />Sta...
References<br />Andrews M. Changing markets, changing relationships. Library Hi Tech. 2007;25(4):562-78.<br />Antelman K, ...
Acknowledgments<br />Many thanks to my project leaders, Jennifer and Debra, for providing excellent guidance and support t...
Upcoming SlideShare
Loading in …5
×

Discovery Interfaces - The Next Generation

1,776 views
1,689 views

Published on

Abstract:
OBJECTIVE
The objective of this project was to perform an environmental scan of next generation discovery interfaces. This included determining what products and features are currently available as well as exploring NLM’s needs in order to make recommendations for an improved portal to NLM’s varied resources. Currently, the information stored in NLM’s collections can be accessed through several different interfaces, including LocatorPlus (the Voyager OPAC), the NLM Catalog (in NCBI’s Entrez system), the NLM Gateway portal, and, in the future, the Fedora digital repository front page.

METHODS
Over the course of several months, information was collected on different discovery interfaces through a variety of means. Vendor sites, blogs, conferences, gray literature, peer-reviewed articles, and other websites were found through various means, including database searches, conference attendance (e.g., Computers in Libraries), suggestions from the project leaders, by monitoring listservs (code4lib and NGC4LIB), and by “berry-picking” (many of the identified products were uncovered by perusing Marshall Breeding’s Library Technology website, for example). These sources were used to identify specific interfaces, provide examples, and to generalize current practices. Any websites that were deemed useful or potentially useful were collected in a Delicious (an online bookmarking website) account. Article citations found in library literature databases were collected using EndNote Web. In addition, the author identified and interviewed key players both inside and outside the National Library of Medicine with the help of the project leader. Trends were identified by compiling the interviews. Information was also gathered on the National Library of Medicine’s current catalog interfaces informally through NLM’s fact sheets and the interfaces themselves.

RESULTS
At final count, the scan resulted in a bibliography of 15 items, a recommended reading list of 75+ items, 6 internal interviews, and 3 external interviews. In addition, about 80 websites have been collected in Delicious, including example library catalogs, platform documentation websites, and related blog posts. This collection of websites may continue to grow as future work on an NLM discovery interface progresses. A comparison of features and which discovery platforms include them also resulted from this scan. The most pervasive theme found throughout is the call to design library catalogs that are intuitive to users; i.e. that work similarly to other commonly used web discovery interfaces (Google, Amazon.com, etc.). Associated with that theme is the request for features such as spell-checking, the inclusion of book covers, faceted search results, and the idea of a “one-stop shop.”

CONCLUSIONS
The findings of this scan support the idea that the National Library of Medicine can and should move towards improving its discovery interface in order to aid users in their quest for information and to expose NLM’s unique collections. There are numerous options available, both from commercial vendors and community-supported open source projects. NLM, with its strong technical capabilities and staff strengths, may do well to explore an open-source solution that can be manipulated in house. However, the key to making a decision will be to solidify which features are the most important and to do a cost-benefit analysis. User studies should also be encouraged, especially since NLM’s users are difficult to define. Finally, NLM should work to stay informed of trends and to be actively involved in discussions on discovery interfaces.

Published in: Technology, Education
1 Comment
2 Likes
Statistics
Notes
  • The notes are here, but please realize that occasionally it's an exact script, and occasionally it's just cues. :)

    Slide 24 has animation for the mockup. You can download the presentation if you just want to look at the individual images, or you can contact me if you'd like the actual animation.
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total views
1,776
On SlideShare
0
From Embeds
0
Number of Embeds
18
Actions
Shares
0
Downloads
35
Comments
1
Likes
2
Embeds 0
No embeds

No notes for slide
  • The motivations for this project are related to the objectives of the Collection, Preservation, and Access to Information Strategic Planning working group. JM: NLM currently provides multiple interfaces to its many silos of data - traditional bibliographic catalog , full-text documents and digital collections. LO’s goal is to eventually provide fewer or one comprehensive discovery platform to its collections that offers a wide range of discovery and delivery experiences.It was not a goal to evaluate the current systems.
  • First, I wanted to distinguish between front and back ends. The scope of my project was to look solely at front ends, or the public interface, often called the OPAC, which can be considered separately from the cataloging, acquisitions, circulation, etc. modules. I also wanted to give a little background on public access catalogs at NLM. The National Library of Medicine’s catalog started with the printed Index Catalogue in 1880 (which continued in series until 1961), moved to the card and microform catalogs, and finally switched to online access only in 1981.CITE (“Current Information Transfer in English,” developed in SIS) replaced the card & microform catalogs after a comparison with the “ILS” system from Lister Hill. CITE was a natural-language search-based online catalog. Cite was eventually replaced by another in-house system, Locator, which incorporated Grateful Med. Eventually it was determined that CITE and Locator should be replaced with an out-of-the-box solution that would provide both a user interface and the back-end systems required by staff to process the collections. Brief look at the current systems, search Francis Collins & “neoplasmms”.It works, but it’s not attractive. No spell check.You can’t click through to related items, and what if you were looking for something about Francis Collins? However, “neoplasmms” works a little better.Some things are kind of broken; it’s really hard to see why these results were chosen.
  • I’m going to go over my methodology for collecting the resources for this environmental scan. I found and collected the information in these 3 areas for the purpose ofInterface comparisonUncovering trendsIdentifying key players (people & platforms)
  • For the sake of time, I chose to focus on electronic text that I could access (either freely or through NLM subscriptions) for the majority of my bibliography. I felt this was allowable, especially because this is a fairly recent topic of discussion, and we wanted timely literature about current issues and offerings. I did collect all citations that I thought were relevant in EndNote Web for “recommended reading,” however.
  • The selection of candidates for both internal and external interviewees was established through discussions with the project leaders. Internal candidates were chosen to represent a wide spectrum of positions within LO and all had some form of interaction with the public. External candidates were chosen individually, and represented interface implementer, interface creator, and expert opinion.Kristin Antelman: Associate Director for the Digital Library Administration, NCSUBess Sadler: Chief Architect for the Online Library Environment, UVaMarshall Breeding: Director for Innovative Technologies and Research, VanderbiltInterview questions were modified from Karen Calhoun’s report: “The Changing Nature of the Catalog and its Integration with other Discovery Tools.” This report was also looking for trends, so we chose to use the questions as a loose base and tweaked them to match our specific interests in NLM needs and priorities. [print out list of interview questions to pass around]Interview procedure (same for everyone): contact via e-mail (with questions), meet in person (internal) or over the phone (external), transcribe notes, send back to interviewee for any revisions or additions. Kristin Antelman was the only exception; she e-mailed her responses back.
  • “Association for Library Collections and Technical Services” – ALCTS (I don’t remember how I heard about this…)Informal. Word of mouth was the main method! Speaking with project leaders, and others (thanks John!), who originally pointed me to sites such as Marshall Breeding’s Library Technology site (which led to a lot of berry-picking). I did use search engines (Google) to find home pages for specific products, and occasionally to find example libraries. Additionally, as I have had a personal interest in catalogs, I was already on the first 2 listservs.The websites have been collected on Delicious, and the list can continue to grow. As a side note, Delicious allows for tagging, which I did utilize, but it may not be too useful as it is my own system, and it’s likely I didn’t use it consistently. It was a side experiment, however, and may prove useful in other contexts.
  • -The bibliography is of resources used to inform this project; the reading list includes items such as press releases and older articles related to the topic of discovery interfaces that may be of interest to task forces and working groups looking at this issue.-The collection of websites and the recommended reading list may continue to grow as future work on an NLM discovery interface progresses.
  • The features included here were chosen based on the 3 types of collected resources and in discussions with the project leaders. Two products did not seem to actually be on the market yet: Agent Iluminar and OCLC’s TouchPoint.
  • Throughout my scan, I came across the idea that users expect to be able to use library catalogs the same way that they use the rest of the internet. Additionally, from all of my research, but especially the interviews, it became clear to me that NLM should move…These two points are purposefully general, but they provided the underpinnings for the features I am recommending.
  • -Based on my own impressions from working at the reference desk as well as from the 3 resource areas, I believe any interface NLM uses should have an advanced search. One of our more vocal user groups are librarians, who prefer more powerful searches. -Facets can be thought of as item attributes or dimensions, and any given item can have more than one, such as subject or time period. They group items with similar features. The CITE/ILS study found that NLM users most often search by subject. It also helps users apply limits as they look at search results, rather than before. -Visuals. Although book covers may not be useful for our users in choosing one book over another, they are helpful on an interface screen as visual cues and break up long lists.-Pages with no results should disappear. The possibility exists to leverage our own powerful UMLS to help.-NLM has a number of unique collections that should all be found in one place (results should be able to indicate which collection an item was in originally, and facets will help limit to only one collection). The NLM Gateway allows users to search across collections, but it does not seamlessly integrate results.--These features are becoming standard in the new interfaces that are coming out, although not every interface has all of these. However, both simple search & relevancy ranking are in fact in every single interface I explored. So while I felt they were certainly important (and the literature/interviews/web sites support this), they are no longer clear factors in choosing one interface over another.However, the last two are not standard:-The need for a browse feature should be explored. The ability to simply scroll through everything around a given starting point is a way that users discover information and should be considered.-Access to everything possible is one of those features internet users have come to expect. Link resolving provides the ability to push users out to resources such as Hathi Trust, Google Books, etc. could be extraordinarily helpful, perhaps especially so for users outside the Reading Room (where link resolving also serves to take users to subscription content).JM: FYI we are undertaking a project in the Fall to harness open URL link resolving (starting w/ SFX)
  • Here’s some information on my recommended interfaces, including example links and lists of features. I pulled out my recommended features and determined whether or not each interface has them. I apologize for the small print!My recommendation is for NLM to go with an open-source product. I think that the in-house skills exist here, and that these solutions will give NLM the flexibility it needs as such a unique setting. There are several options, including OpenBib and Scriblio, but Blacklight & VuFind look like they may be the best options for NLM. My preliminary impressions are that VuFind has strengths in its APIs and its advanced search feature, while Blacklight is specifically designed to include non-book records and has developed customizable interfaces for different user groups. Blacklight also already works with Fedora repositories. Both are built on SOLR indexes (open source enterprise search server), have strong communities that support and improve the products and both incorporate numerous features, including most of my recommended features. Most importantly, I think either of these products would serve NLM users very well, and would allow NLM to customize and to be involved in an open source project that may benefit many different groups, including medical libraries who follow in NLM’s footsteps.However, if it is determined that NLM would be better served by a proprietary service, my current recommendation is for Ex Libris’s Primo. I choose Primo for several reasons. First, I eliminated several products (LS2 PAC, SirsiDynix’s Enterprise, etc.) because they appeared to be more suited for public libraries with more traditional collections. It should also integrate seamlessly with Voyager, another Ex Libris product (although Primo can also work on top of other ILSs, should NLM decide to switch backends in the future). It also incorporated all of my main recommended features, and addressed further considerations such as accessibility and having an open API. Aquabrowser and the Endeca platform had similar features and strengths and should also be considered, but based on the implementations I saw, Primo seemed the best fit for NLM.
  • But I want to demonstrate that things are constantly changing. It’s possible there will be new interfaces or improvements to interfaces I eliminated that would make them appropriate for NLM.
  • These other considerations should also be addressed when NLM evaluates its options.-Issues regarding 508 compliance, especially given the number of Java-created features.-Content visibility includes harvesting by search engines, and also creating tools like RSS feeds and having permanent URLS. Marshall Breeding pointed out that “the goal shouldn’t be for library content to be found solely through search engines. We should be developing our interfaces so that once users come to us, through whatever means, they recognize our value and come directly to us and stay when they have information needs.”-Having an API (Application Program Interface) for the discovery interface is important and could have a lot of potential. In addition to providing “a way for others to help” us, an API allows us to create necessary applications in house.
  • Some other cool examples, if there’s time1st two, in house examples.Dartmouth: 1 huge index (search vladimirnabokov), although it has its problems.American University: cool things that can be done on top of some systems.
  • Finally, I wanted to include some recommended general actions.The first is evaluate:-Numerous interface options are available – both from commercial vendors and community-supported open source projects.A cost-benefit analysis will be the key to making a decision to solidify which features are the most important. Cost in staff time and money should be included.-Evaluate what users want, and possibly who they are, to help with any decision making regarding the interface and its features. Making the decision to reach out to users beyond the reading room (or not) will also help. - Finally, evaluation from the outset of any implementations will help drive improvement and see what’s working and what’s not. In other words, have an evaluation plan in place even before implementing a pilot.Secondly, monitor the listservs & blogs, keep attending conferences and reporting back.Example, the National Agricultural Library has sponsored a local code4lib meetup. Could also work with open source communities. Sharing experiences by publishing evaluation results is another example.NLM has a lot of talent it could leverage (from in-depth coding to those with eyes for good ideas). Allowing people to play with the possibilities will create stronger systems and may also help recruit & retain more talent.
  • I wanted to end with a little thought experiment: combining many of the cool features that are out there (and specifically incorporating all of my recommended features), I created this mockup version of a discovery interface for NLM.
  • Discovery Interfaces - The Next Generation

    1. 1. Discovery Interfaces: The Next Generation<br />To boldly go where nOPAC has gone before…<br />Amy Donahue<br />Project Leaders: <br />Jennifer Marill & Debra Thangarajah<br />
    2. 2. Outline<br />Objective<br />Motivations & Goals<br />Introduction<br />Methodology<br />Literature<br />Interview Results<br />Websites<br />Results<br />Discussion & Recommendations<br />Features<br />Interfaces<br />Actions<br />Mock-up<br />This report is a requirement for the completion of the Associate Fellowship Program at the National Library of Medicine. The views and opinions of the report author do not necessarily state or reflect those of the U.S. Government, and they may not be used for advertising or product endorsement purposes. <br />
    3. 3. Objective<br />To perform an environmental scan of next generation discovery interfaces. <br />
    4. 4. Motivations & Goals<br />LO Strategic Plan on Collection, Preservation, and Access to Information<br />Update older systems<br />Incorporate new features<br />Provide a comprehensive interface to LO collections<br />Determine what products and features are currently available<br />Explore NLM’s needs in order to make recommendations<br />
    5. 5. An Introduction to Catalog Interfaces<br />Front end vs. Back end<br />Public interface / ”librarian” modules<br />“We should no longer speak of the ‘ILS’ as a stand-alone product (if, in fact it was ever a truly stand-alone product).” Mark Andrews<br />History & background <br />Index-Catalogue <br />CATLINE interfaces (CITE, Locator)<br />Current NLM interfaces<br />LocatorPlus (Voyager)<br />NLM Catalog<br />NLM Gateway<br />
    6. 6. Methodology<br /><ul><li>Literature
    7. 7. Interviews
    8. 8. Web </li></li></ul><li>Literature<br />Finding:<br />Library database search: <br />Ovid – Wilson Library Literature<br />EBSCOhost – Library, Information Science and Technology Abstracts<br />Article bibliographies <br />Collection:<br />Endnote Web<br />
    9. 9. Interviews<br />Selection of candidates (in order of interview)<br />NLM: Wanda Whitney, Dee Clarkin, John Doyle, Dianne McCutcheon, Joyce Backus & Martha Fishel, Laurie Duquette<br />External: Kristin Antelman, Bess Sadler, Marshall Breeding<br />6 Interview questions, including<br />How could a new NLM catalog/discovery interface maximize its utility for the communities and users served by NLM? Please also describe the specific communities and users you are thinking about.<br />What specific features would you like to see included in a new interface?<br />Interview procedure<br />
    10. 10. Websites<br />Finding:<br />Word of mouth<br />Conferences<br />“Berry-picking” (linking to different sites from original)<br />Listservs:<br />code4lib<br />NGC4LIB<br />ALCTS forum<br />Collection:<br />Delicious.com<br />
    11. 11. Results<br />
    12. 12. Results<br />Collection of 105 resources<br />Bibliography of 16 resources<br />Recommended reading list of 99 items<br />6 internal interviews and 3 external interviews<br />~80 websites have been collected in Delicious <br />A comparison of features and which discovery platforms include them <br />
    13. 13.
    14. 14. Discussion &Recommendations<br />
    15. 15. Discussion points<br />The most pervasive theme found during the scan was the call to design library interfaces that are intuitive to users.<br />The National Library of Medicine should move towards improving its discovery interface in order to aid users in their quest for information and to expose NLM’s unique collections.<br />
    16. 16. Recommended Features<br />Advanced search<br />Facets<br />Visuals<br />“Did you mean”/spell check<br />Inclusion of multiple sources<br />Simple search/relevancy ranking<br />Browse<br />Incorporation of SFX/Umlaut link resolving to take users to outside resources<br />
    17. 17. Recommended Interfaces<br />
    18. 18.
    19. 19. Other Interface Considerations<br />Accessibility (508)<br />Mobile interfaces <br />Language<br />Content visibility<br />APIs<br />Brooklyn Museum<br />An API “…provides a way for others to help you…These APIs allow for various information aggregation and/or discovery services to be created on your behalf.” Roy Tennant, TechEssence.info<br />
    20. 20. Other Examples<br />World Digital Library<br />National Library of Australia Prototype<br />Dartmouth: Summon<br />American University of Rome<br />
    21. 21. Recommended Actions<br />Evaluate<br />Interface options (cost-benefit analysis)<br />Users<br />Chosen interface<br />Stay informed<br />Become involved<br />Utilize technological skills<br />
    22. 22.
    23. 23. References<br />Andrews M. Changing markets, changing relationships. Library Hi Tech. 2007;25(4):562-78.<br />Antelman K, Lynema E, Pace AK. Toward a Twenty-First Century Library Catalog. Information Technology and Libraries. 2006;25:128-39.<br />Tennant R. TechEssence.info [blog on the Internet]. [Sonoma (CA)]: Roy Tennant. [2005 Nov] - . If it doesn&apos;t have an API, it&apos;s not worth having; 2006 Dec 13 [cited 2009 Jul 14]; [about 2 screens]. Available from: http://techessence.info/node/81<br />Picture credits: University of Australia, NguyenDai, HMD Images, University of Michigan, University of Minnesota, NCSU, UNC-Chapel Hill<br />
    24. 24. Acknowledgments<br />Many thanks to my project leaders, Jennifer and Debra, for providing excellent guidance and support through the project.<br />Also many thanks to those who allowed me to interview them: they provided so much more than their time. Wanda, Dee, John, Joyce, Martha, Dianne, Laurie, Kristin, Bess, & Marshall. <br />Joyce, again, for her role as my preceptor. <br />Thanks also go to Kathel Dunn, Jason Broadway, and Sally Sinn and my fellow Fellows.<br />Finally, I’d like to thank NLM, and especially Dr. Lindberg, Betsy Humphreys, Sheldon Kotzin, and Becky Lyon for their support of the AFP.<br />
    25. 25. Questions?<br />

    ×