I am a Ph.D. candidate in the School of Computer Science at the University of Massachusetts Amherst. I work with Professor Leon Osterweil in the Laboratory for Advanced Software Engineering Research (LASER).

My work is the area of process definition and improvement. More specifically, I have developed an approach that can be used to model and accommodate process and system variation in a formal and analyzable way. I am interested in how different techniques for accommodating variation inform desiderata for process definition languages. I have concentrated on two different case studies: one on providing process guidance in conflict-resolution and negotiations, and another one on analyzing variation to ensure trustworthy elections. In summer 2010, as a visiting researcher at Lero - the Irish Software Engineering Research Centre, I worked with Professor Bashar Nuseibeh on applying process modeling and analysis techniques to security requirements definition and argumentation.

The first case study was done in collaboration with the National Mediation Board, a federal agency that provides mediation and arbitration services. Our purpose was to precisely capture the way the NMB conducts mediation in a face-to-face settings and then translate that into a process definition in the Little-JIL process modeling language that can be used to guide negotiation in a remote, distributed setting on the internet. This process definition can provide explicit process guidance to support mediators in the field and train new mediators. To achieve this goal, we developed a system for online dispute resolution, called STORM2, which provides explicit process guidance; you can see a demo here. What is interesting about the process definition is that as we strive to capture the process accurately and in sufficient detail, we find that mediators often develop their own styles and vary the exact details of the process at lower levels of abstraction. As a result, different process variants must be created because a single process definition cannot precisely model these different "styles." Although each variant is a new process, these processes are related according to prescribed variation relationships so it may be more beneficial to consider them as a family of process definitions (see "Generation, Composition, and Verification of Families of Human-Intensive Systems," "Characterizing process variation (NIER Track)," "Categorizing and modeling variation in families of systems: a position paper," and "Representing Process Variation with a Process Family").

The second case study concentrates on modeling variation observed in real-world election processes to generate families of systems, and then analyzing these systems for vulnerabilities. There are several different analysis techniques that could be applied;  for example, we can precisely define formal requirements and then use finite-state verification to check if all variants in the family adhere to these requirements. What is interesting in this domain is that requirements encompass sometimes conflicting interests--for example, juxtaposing the privacy of the voter against the assurance that no voter casts more than one vote. This issue can easily get very complex when provisional ballots are introduced. Another, complementary technique we can use is fault tree analysis, where we examine how the misperformance of a single step may result in disastrous consequences. These analyses, along with others, can be applied to the family of systems in order to identify improvements in the real-world process. For more information, see "Modeling and analyzing faults to improve election process robustness" and "Specifying and Verifying Requirements for Election Processes." This work is being continued under an NSF CISE cross-cutting grant and a National Institute of Standards and Technology (NIST) award in collaboration with Professor Matt Bishop and his lab at UC Davis.


Contact information:

Borislava I. Simidchieva

316 Computer Science Building

140 Governors Drive

UMass Amherst

Amherst, MA 01002

bis at cs dot umass dot edu