The Amateur Movie Database: Archives, Publics, Digital Platforms

I have a confession to make: I am a digital humanities (DH) amateur. After reading about new digital approaches to film scholarship and developing my own project, I quickly discovered that, while imagining a DH project is an exciting proposition, building one is another thing altogether.¹ The Amateur Movie Database (AMDB) began with a rather straightforward goal in mind: to collect information about significant amateur films in one place so that researchers, archivists, and the interested public can investigate this category of filmmaking. Now the AMDB (http://www.amateurcinema.org/) faces a range of challenges around refining what the information collected should contain, specifying which groups are most likely to use the database and how, and selecting the most appropriate digital tools for the task from the wide array of available DH tools and methods.

In terms of refining the information our project might collect, Acland and Hoyt's recent Arclight Guidebook to Media History and the Digital Humanities provides a valuable introduction to the broad categories of DH projects. These include ones that transform information into data sets, digitize materials for wider dissemination (typically via the internet), use digital technologies for new kinds of multimedia publication and interactive presentation, and employ social media and networked communities to help build a DH project in progress.² The AMDB project employs—or could potentially employ—each of these elements to some degree.

The goals of the AMDB are closely aligned with the first kind of DH project, that which transforms information into a new data set. We begin with a selection of significant North American amateur films produced between 1928 and 1971. This corpus includes winners of amateur movie contests as well as films that are determined by project collaborators to be otherwise significant (having gained scholarly or archival recognition, such as National Film

Preservation Foundation preservation grants, or representing works from marginalized communities). This corpus consists of approximately thirteen hundred films, including relatively well known works like The Fall of the House of Usher (Watson and Webber, 1928) and Multiple Sidosis (Laverents, 1970). But there are many more unheralded films. These film data are already a productive set for analysis, but there is still room for additional future contributions, such as winners of other movie contests or films that are discovered in archives. An area for future expansion is the geographical range of the project, widening the focus from North American filmmakers to include more information about international films.

The AMDB is unlike typical library or archival databases in that the films included are not ones that we possess. In fact, one of the goals of the project is to compile a large list to use for locating extant films and evaluating the proportion of films that have been lost. In this sense, the project is like the AFI Catalog of Feature Films and the Library of Congress's (LOC) American Silent Feature Film Database, which include both extant and lost films.³ The AMDB resembles a catalog in that it compiles information about the films (and filmmakers) into discrete fields, making the data more "granular" and available for metadata searching and analysis. Like the AFI resource, the AMDB mines textual documentation to provide content summaries and production details. Like the LOC database, we are interested in identifying which films have survived, and where.

The Amateur Movie Database was designed (by our technical guru Jack Brighton) with PBCore metadata standards in mind so that the data could be shareable in the future. The data that we collect come from a variety of sources, including textual documentation in published sources (such as amateur movie magazines), archives, online sources (especially the Media History Digital Library and the Internet Archive), and (when possible) the films themselves. The challenge here is to find and enter pertinent information about the films, a task that progressively is more research intensive. After basic information (title, filmmaker, and year) is entered, finding more information about the films (descriptions, images, locations, etc.) requires item-by-item research.

Figure 1.

AMDB record for Fairy Princess (Margaret Conneely, 1956...