The ACM has just launched a new version of its Digital Library (DL), the first major overhaul of its vast store of computing literature in almost 10 years. The information services offered by the new library have been enhanced and tweaked dramatically, usability and performance have been improved, connections to external services have been added, and much new content, including multimedia, is available. Most fundamentally, the DL has evolved from a relatively simple keyword search and journal article retrieval model to one in which users can see the connectionsand the strength of connectionsamong topics, journals, articles, authors, collaborators, institutions, Special Interest Groups (SIGs), and conferences.
"It's all about showing users the context in which something fits, so they know that there is more to this space than just their ability to grab an article and go," says ACM's Director of Information Services Wayne Graves. "It's about surfacing and leveraging information in different ways."
The DL now explicitly recognizes that many users do not work top-down from the library's home page (the original model) but come in directly at the citation page for a given articleoften sent there by an external search engine. "The citation page has become the front door," explains Graves, the chief architect of the new library system. "So a design goal was to get as much information there as we can."
Information on the citation page now appears in three logical blocks. The first contains the basic information about the article such as title, author, and publication name; links to author profiles; links to the journal or conference home page, table of contents, and archives; and bibliometrics (numbers of downloads and citations). The second block serves up clickable options to buy an article, to request permissions, to save the article to a personal binder or workspace, to export it in various formats, or to send it via email or to external services such as Slashdot, Facebook, and Twitter. The third section has 10 tabs that bring up information about the article, such as source materials, references, citations, reviews, and comments. An "index terms" tab shows pointers to other articles on the same and related topics.
Joshua Hailpern, a Ph.D. candidate in computer science at the University of Illinois at Urbana-Champaign, has used ACM's digital services for nearly a decade, and recently served on a panel of users to evaluate ACM DL prototypes. "The information used to be sort of dumped on the user in this long, difficult-to-navigate layout," he says. "What they did is ask, 'What's the information users care about first?' And they kept it simple."
A key enabler of many of the DL's new capabilities, and for a number of older ones, is an architecture that puts a great deal of emphasis on the capture and use of metadatasearching and linkable data about entities in the library, such as publication, author, institution, citation, or SIG. "The amount of metadata that we capture is very large, and that's absolutely key to our new capabilities," says Bernard Rous, director of publications. "We made that as robust as possible so we can do all kinds of calculations and manipulations."
For example, the DL now offers new and powerful ways to look at conferences. Before, a user could retrieve conference proceedings for a past conference, but other information about it, or about future conferences, was not readily at hand. Now a map of the world can be zoomed in on to show ACM conferences in a given city or region. Click on a conference and links to a conference profile page, the conference Web site, and proceedings are revealed. Tag cloud representations provide snapshots of subject area concentration for conferences and SIGswith differing font sizes showing relative importanceand will soon do so for conference acceptance rates over time via color charts.
The new DL reflects user demand for greater interactivity and control of content. For example, a personal binder capability that had served essentially as a way to save search results in a folder has been greatly expanded. An interactive editor has been added so that, for example, a user can annotate a bibliography. "The WYSIWYG editor and HTML forms to comment on articles and to annotate personal binder collections introduces user-contributed content to the DL and allows it to begin functioning as a workspace," Rous says. "It is also far easier to share binders more widely, and the user can generate a single PDF that stitches together all the selected articles and annotations with a cover image. In essence, the new binder function enables the DL to become an authoring and a publishing platform for certain kinds of e-books."
Although ACM already provided one of the most comprehensive global resources for the computing field, a combination of forces prompted this major overhaul of the DL. "User feedback streams in on a daily basis," says Rous. For example, critical feedback from SIGs led to the creation of a Conference Profile page and a SIG Profile page. Analysis of such feedback, augmented by surveys, user workshops, focus groups, and advisory bodies, revealed strong demand for more interactive capabilities and for richer data views.
"Comments relating to change requests are collected, analyzed, and finally placed on a prioritized to-do list," explains Rous. "ACM's Librarian Advisory Group, Digital Library Advisory Board, and Publications Board have all reviewed and critiqued early versions of new functionality. Tests are run with focus groups and, finally, beta production versions are released for further feedback and refinement. The new DL reflects the collective wisdom of a broad community participating in the development process."
The new DL also reflects the reality that many users have dropped their print subscriptions, yet they occasionally want to order a printed copy of an issue. New e-commerce functions provide ways for online-only subscribers to order a print copy and for nonsubscribers to gain access to online content. "A print-on-demand function is sorely needed and will shortly appear along-side the other options," Rous says.
The trend from print to online is also a challenge to external libraries, Rous says, and they have responded by installing complex local infrastructures and by tapping into external systems to manage their digital resources. "ACM is committed to supporting libraries in these transformative efforts," he says.
One way ACM is doing that is by providing access to its increasingly rich metadata, with or without extracted full-texts, for local or third-party indexing to facilitate cross-platform discovery by the library's patrons.
The new ACM Digital Library is more interactive, and puts a lot of emphasis on the capture and use of metadata.
Rous says ACM aims to make the DL the most reliable source for citation and usage statistics in the field, and to make that kind of information useful for the assessment of research and the contributions and influence of individuals and institutions.
Part of the DL's new look is a result of rebranding in response to user confusion over the differences among the terms ACM Digital Library, ACM Guide to Computing Literature, and ACM Portal, with most users simply lumping everything under the "Digital Library" moniker. The full-text library of ACM publications remains, as does the ACM Guide, the larger bibliographic database of computing literature. But these are now more tightly integrated and function as a single resource branded as the ACM Digital Library.
Jill Powell, a librarian at Cornell University's engineering library, says computer science faculty and students have for years used the library in traditional ways"to discover articles by keyword and author"but that usage of it will become more extensive now. She calls the new bibliometrics "very interesting" and says the features for automatic subject and table of contents alerts via email and RSS are attractive. "It's a rich site, it has a lot of features, and it's easy to use," Powell says.
A number of enhancements have been made "under the hood," says Graves, who wrote most of the new code. The pages of the DL are dynamically generated and the data is retrieved from precomputed data sets supported by a new, flatter data model, resulting in a much faster front-end interface. And a new caching scheme fetches only the data needed, rather than the templates and logic required to generate the pages. The library served up about 10 pages per second, on average, under the old architecture but now runs at about 15 pages per second, Graves says. The library contains 1.6 million items, and has users in 190 countries, who download 13 million full-text files and conduct 12 million searches annually.
Rankings of the "influence" of people and institutions, as measured by such data as numbers of publications and their citations and usage history, as well as the frequency and kinds of collaboration, are on the short-term horizon. The new library takes the first small steps toward reporting measures of that influence. The Author Profile page, for example, supplies a snapshot of an author's contribution to the field with coauthors listed as "colleagues." The data that underlies these pages could enable views of professional networks. Indeed, several SIGs are using this data in research projects to produce visualizations of author relationships, Rous says.
Similarly, a beta version of Institutional Profiles aggregates all the information in the bibliographic database to give snapshots of the faculty and graduate students affiliated with that institution at the time of their publications. Their output is weighted and their publications are listed. The subject areas of greatest concentration at the institution are visually displayed. These profiles reveal for the first time the frequency of author collaborations across different institutions.
"This to me is the really exciting direction of the DL," Hailpern says. "There is amazing stuff that can be done with information networksnot just saying here's the word-count frequencies in a tag cloud, but really saying, 'How influential is this paper, what are the terms and areas and authors this paper has influenced? What is the world it relates to?' This is the kind of stuff that ACM is uniquely positioned to leverage because no other community is so in touch with these information networks."
©2011 ACM 0001-0782/11/0200 $10.00
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from firstname.lastname@example.org or fax (212) 869-0481.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2011 ACM, Inc.
No entries found