UNIVERSITY OF CALIFORNIA, DAVIS

The cryptographic community has a very particular disciplinary culture, and this culture shapes the set of problems that tend to be noticed and the style of solutions that tend to emerge. This is of course true in any scientific field, but, with cryptography, the influence of cultural norms seems particularly pronounced. Assumptions underlying the field's disciplinary culture include beliefs that cryptographic papers are like papers in mathematics, equally clear, trustworthy, and precise; that cryptographic definitions are refined, stable, and usable; that cryptographic work is strongly allied with computer security, and plays a major role in constructing secure and trustworthy systems; and that protocols, proofs, and attacks are the cryptographic community's outputs of most lasting significance. In this project, we are working to identify and challenge such culturally-rooted assumption can lead to valuable work, including the recognition of important problems that would likely not be seen from a vantage point more accepting of the status quo. The problems the PI intends to address include: * Developing a useful theory for partially specified protocols, starting with entity authentication, to enable the provable-security treatment of real-world protocols without eliding significant aspects of them first. * Creating code-based definitions for cryptographic goals whose conventional exposition would involve pages of English prose. * Designing practical cryptographic hash-functions out of permutations--that is, fixed key blockciphers--and employing sophisticated computer-based methods to analyze the underlying compression functions. * Embarking on a program to bring the best-possible authenticated-encryption schemes into popular use, the program involving the creation of algorithms, publicly-available code, timing studies, APIs, and RFCs. * Engaging in empirical studies on the sociology of the cryptographic community, these including bibliometric studies and literature surveys aimed to better understand the community's character. * Defining and exploring zero knowledge from a framework of concrete (not asymptotic) security. * Refining a course in ethics and technology and engaging in scholarship and outreach to try to popularize a broadened view of what should be the focus in ethics courses for computer scientists and engineers. While the motivation behind this project is somewhat contrarian, the proposed work encompasses the kind of activities that are well-regarded in the field: developing compelling definitions, constructing low-level primitives, finding novel ways to prove them secure, and analyzing emerging security standards. Papers produced under this project will meet the highest intellectual standards and will be appropriate for the top venues in the field. Some of the work pushes, in a healthy direction, the conventional understanding of what kind of work has high intellectual merit. This proposal was born out of dissatisfaction over the broader impacts of the PI's field. It aims to push cryptographic research in a direction that will increase the utility of its models, its connectedness to computer security, and the accessibility of cryptographic definitions. The proposal is unusually rooted in historical and cultural considerations about the disciplinary character of a field. The work is grounded in a belief that inessential aspects of the collective identity of the cryptographic community have attenuated the broader impacts of the field, but that this outcome is not a legacy set in stone.

Clarification of Codes