User Research To Completed Product: OVDLT

Developing the Open Video Digital Library Toolkit (2005-2009)

Project Motivation

The Open Video Project, which I established with my doctoral studies advisor in 1999, was one of the first substantial collections of publicly accessible digital video available on the World Wide Web. After the site's creation, I received many requests for the site's programming code and database from organizations that possessed video collections but lacked the technical expertise and funds to develop their own digital video library. Motivated by the clear need I saw, I wrote a grant proposal and received funding to develop an open source digital video toolkit (hereafter, "Toolkit") that would provide a no cost solution for these libraries, archives, museums, and other organizations.

User Needs Assessment

Although I had a pretty good idea of what people were looking for in digital video library software (mostly based on email inquires about the Open Video project, but also from my understanding of the research developments in the area and informal discussions with potential users), this was anecdotal evidence. Before embarking on development of the Toolkit it was important to gain a broader-based understanding of the characteristics of organizations' video collections, their technical capacity, and the goals and search preferences of their intended end-users. Using a web-based survey (n=83) and followup one-on-one interviews (n=21), I obtained a rich set of data that gave me a much more solid idea of what would be most useful in the Toolkit.


The survey and interview data provided the basis for developing two sets of user personas (imaginary but representative potential users of the Toolkit, which help convey the needs of the target audience and prioritize requirements). One set of personas represented the people at organizations that would implement the Toolkit product; that is, those with the video collections they want to share with their patrons. The second set of personas represented these patrons, or end-users of the organizations' Toolkit-based digital video collections. Each set of four personas attempted to capture and distill the characteristics and needs evident in the user needs assessment data.

Concept Models

The user needs assessment data also enabled me to develop several concept models. These concept models helped provide the "big picture" of the Toolkit, useful both for myself and others involved in developing the product, and for communicating the goals of the project with those outside the development team. The first concept diagram at left helps explain how the Toolkit would be used by an organization to catalog, publish, and maintain their digital video library. The next image (available in the popup) is a high-level conceptual view of the main activities expected of end-users of a Toolkit-generated digital library.

Site Map

Moving towards more specificity, I developed a site map that incorporates the user needs assessment data and the direction suggested by the concept models and offers a somewhat more concrete representation of intended product. The site map conveys both hierarchical and navigational relationships among the primary features of the product (the image at left represents only the end-user aspect of the product, not the administrative/cataloging side). This diagram helps show how features fit together, and provides a basis for making things even more concrete.


Every element shown in the site map was next fleshed out in a wireframe diagram. The wireframes clearly represent how user requirements derived from the user needs assessment studies will be met on the site. Each wireframe represents one page of the intended end-product and serves as a concrete blueprint for the programming and design stages of development. Many minor aspects of the wireframes were modified during the programming and design stages, but for the most part the wireframes fairly closely reflect the completed Toolkit product.


I hired a contract programmer to do most of the programming work on the Toolkit, although I also did some programming myself. With the wireframes as the primary guide, I used the web-based Pivotal Tracker tool to create specific programming requirements (shown in the image at left); the contract programmer and I could then use Pivotal Tracker to discuss, modify, prioritize, and check-off the programming requirements when completed. The Toolkit is written in Ruby on Rails, with MySQL as the database. We used Git for version control and GitHub for the project code repository.

Visual Design

When the programming code for the Toolkit was largely completed, I turned to the visual design of the product. I contracted out most of the visual design, though again I also did some of this work myself. We worked through several stages of visual design mockups (the first image at left is an early version) but the final visual themes are fairly close to the layout suggested by the wireframes. There are currently two distinct visual themes available for the Toolkit (the image below left is an example of the second theme).


Testing of the Toolkit took place at many stages of the product development. As features were developed and parts of the product became functional, a student assistant would do informal usability testing and provide formative feedback. At a couple of different stages of development, our project partner organization would test the product with their own video collection and provide more formal feedback. Finally, in 2010-2011, a faculty colleague and I used the Toolkit for student projects in two courses (image at left is one example) and in a research study; these uses provided additional formative feedback and ideas for new features.

Product Release

The completed Toolkit product is completely free and open source, and runs on Linux and Mac OS X systems that have the appropriate software prerequisites installed. It is available for download at GitHub. Sample installations include the "Classic Cartoon Collection" and "Northeast Historic Film's Moving Things" (these servers are no longer available).

(Note that this is no longer an active project because I was not able to continue work on it after changing jobs in 2011.)