Metavid.org: A Social Website and Open Archive of Congressional Video Michael Dale, Abram Stern, Mark Deckert, Warren Sack University of California Communications Building, Room 101, 1156 High Street Santa Cruz, CA 95064 {dale,aphid,mdeckert,wsack}@ucsc.edu Abstract We have developed Metavid.org, a site that archives video footage of the U.S. Senate and House floor proceedings. Visitors can search for who said what when and also download, remix, blog, edit, discuss, and annotate transcripts and semantic metadata. The site has been built with Open Source Software (OSS) and the video is archived in an OSS codec (Ogg Theora). We present a walk through of the architecture of Metavid and highlight two aspects of its design: (1) open standards; and, (2) Wiki functionality. First, open standards allow Metavid to function both as a platform, on top of which other sites can be built, and as a resource for “mashing.” For example, Voterwatch.org pulls its video from the Metavid archive. Second, Metavid extends the MediaWiki software (which is the foundation of Wikipedia) into the domain of collaborative video authoring. This extension allows closed-captioned text or video sequences to be collectively edited. Introduction Metavid.org site hosts an archive of video footage of the U.S. Senate and House floor proceedings indexed so that one can search for who said what when. Visitors can download, remix, blog, edit, discuss, and annotate transcripts and metadata. The site has been built with Open Source Software (OSS) and the video archived with OSS codecs such as (Ogg Theora). Incorporating open standards and Wiki functionality, the Metavid architecture facilitates deliberation, consensus building, and the free sharing of ideas. We have built five different interfaces for Metavid. Each interface corresponds to a different use or possible set of users of the Metavid archive: (1) citizen interface; (2) blogger interface; (3) media maker (e.g., film maker) interface; (4) API, application program interface; and, (5) platform developer interface. The citizen interface allows the casual visitor to search for and play video clips in a web browser. The blogger interface allows video clips from the Metavid archive to be linked and embedded in Copyright © 2008, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. other web pages and standard blogging platforms (e.g., blogger.com). The media maker interface provides the means for clips from the archive to be edited together into sequences in a browser-based tool; moreover, all the clips can be downloaded in high-resolution formats. The Metavid API gives application developers the means to query, sequence, download and mash-up the resources of the archive programmatically. And, finally, the platform developer can download the entire code base of the project and repurpose it for another archive. For example, the meetings of the City Council of Missoula are now being archived using a version of Metavid. In short, Metavid is a platform that provides a number of different interfaces to its archive thus making it possible for others to browse the archive or build on top of it by “re-skinning” it, “mashing” it up with the contents of other sites, or repurposing its content or code for entirely new projects. The Archive Metavid can be found at http://metavid.org. It is a system of servers and Open Source Software that currently captures, archives and streams U.S. House of Representatives and Senatorial floor proceedings. The archive begins in January 2006 and extends to the present. These proceedings are produced by governmental employees and fed to various media outlets. Our archive, thus far, consists of C-SPAN's rebroadcast of this public domain video. We also capture the closed-caption stream (as produced by the National Captions Institute in cooperation with Congress) and perform automatic character recognition on the graphics broadcast with the video. Consequently, using Metavid, one can search the archives for clips of specific Senators or Representatives addressing specific topics. The Metavid servers run on the Fedora GNU/Linux operating system. Three of the computers are used to operate the website and serve media content and one is dedicated to video and closed-caption digitization via a cable feed and television tuner cards. Using two Hauppauge PVR-150 tuner cards and Linux drivers from the IVTV project, the video stream is initially captured in