contractor teams to develop specific risk management approaches that are appropriate to a project's particular technology objectives. Within the agency itself, those approaches are codified in an internal pragmatic process using technical criteria milestones that effectively constitute a contract between the technical program manager and the agency director for the execution of the work.
To ultimately transition the products of technology to armed services customers, DARPA must be able to connect the applicability of products from its research to emergent military needs. In the DOD this applicability is usually communicated using formally defined sets of requirements and key product definitions. For the most mature capabilities, DARPA will execute a memorandum of agreement with the responsible service organizations defining those organizations’ roles and responsibilities for technology maturation, prototype demonstration, and transition to acquisition. In doing so, DARPA makes extensive use of the NASA technology readiness level (TRL) system to communicate the estimated maturity of products to DARPA customers. The TRL scale is used largely because it has been widely adopted by the larger military acquisition community.
Welby also stated that DARPA program managers had an ambivalent relationship with what the military calls “requirements.” Requirements are a collection of specifications, features, and capabilities that defines an approved military materiel product available for acquisition or purchase. While program managers understand that requirements capture important elements of a product's function, Welby said program managers often perceived them as limiting innovation, producing inefficiencies, and reflecting the status quo because they were often defined in terms of the presumed product rather than the statement of need. As a result, DARPA often stretches, ignores, bypasses, or seeks to change requirements that exist at the outset of programs. This is part of the project-specific relationships that a program manager develops in executing an effort. Service requirements tend to influence what DARPA management decides to fund, but they are not the only factor in the decision.
For a number of years, DARPA has asked its program managers informally to answer several questions before embarking on any program:
What is this project trying to do?
How is it done today? What are the limitations of the current process?
What is truly novel about the approach being suggested that will remove those limitations and improve performance? By how much will performance improve?
If successful, what difference will the project make?
What are the interim technical milestones required to prove the hypothesis?
What is the transition strategy?
How much will it cost?
Are the programmatic details clear?
Welby stressed that DARPA depended on an extremely energetic set of very bright program managers to answer those questions and to arrive at solutions that offer transformative impacts. Often, this is done with the help of military end users. However, the solutions developed are rarely under the control of those users, at least while the effort is funded by DARPA.