Integrating User-Generated Content in the ACL Anthology

Proceedings of the ACL-2012 Special Workshop on Rediscovering 50 Years of Discoveries, pages 83–87, Jeju, Republic of Korea, 10 July 2012. c⃝2012 Association for Computational Linguistics Integrating User-Generated Content in the ACL Anthology Praveen Bysani Web IR / NLP Group (WING) National University of Singapore 13 Computing Link, Singapore 117590 bpraveen@comp.nus.edu.sg Min-Yen Kan Web IR / NLP Group (WING) National University of Singapore 13 Computing Link, Singapore 117590 kanmy@comp.nus.edu.sg Abstract The ACL Anthology was revamped in 2012 to its second major version, encompassing faceted navigation, social media use, as well as author- and reader-generated content and comments on published work as part of the re- vised frontend user interface. At the backend, the Anthology was updated to incorporate its publication records into a database. We de- scribe the ACL Anthology’s previous legacy, redesign and revamp process and technolo- gies, and its resulting functionality. 1 Introduction To most of its users, the ACL Anthology1 is a useful open-access repository of scholarly articles on the topics of computational linguistics and natural lan- guage processing. The liberal use and access policy granted by the Association of Computational Lin- guistics (ACL) to the authors of works published by the ACL makes discovery, access, and use of its re- search results easily available to both members and the general readership. The ACL Anthology initia- tive has contributed to the success of this mission, both as an archiving and dissemination vehicle for published works. Started as a means to collect and preserve arti- cles published by the ACL in 2001, the Anthology has since matured and now has well-defined work- flows for its core missions. In 2009, the Anthology Praveen Bysani’s work was supported from the National Research Foundations grant no. R-252-000-325-279. 1http://aclweb.org/anthology/; beta version 2 currently at http://aclantho3.herokuapp.com/. staff embarked to expand the Anthology’s mission to meet two specific goals: on the backend, to enforce a proper data model onto the publication metadata; on the frontend, to expand the scope of the Anthol- ogy to encompass services that would best serve its constituents. Where possible, we adopted widely- deployed open source software, customizing it for the Anthology where needed. With respect to the backend, the revamp adopted a database model to describe the publication metadata, implemented using MySQL. On top of this database layer, we chose Ruby on Rails as the application framework to interact with the data, and built suit- able web interfaces to support both administrative and end-users. The backend also needed to support resource discovery by automated agents, and meta- data export to sites that ingest ACL metadata. With respect to the frontend, the Anthology web- site needed to meet the rising expectations in search and discovery of documents both by content and by fielded metadata. To satisfy both, we incorporated a faceted browsing interface that exposes metadata facets to the user. These metadata fields can be used to restrict subsequent browsing and searching actions to the values specified (e.g., Year = 2001– 2011). Aside from resource discovery, the fron- tend also …

Content truncated to 500 words