Computer Science Colloquia
Friday, January 18, 2013
Advisor: John Knight
Attending Faculty: Jack Davidson (Chair), Dave Evans, Alf Weaver, and John Lach (Minor Representative)
1:00 PM, Rice Hall, Rm. 242
Ph.D. Project Proposal Presentation
Towards Establishing Desired Properties in Software of Unknown Provenance
The proposed research investigates the concerns of users about software
for which minimal or no software development information or artifacts
are available, in particular the software source code. This software,
referred to as Software Of Unknown Provenance (SOUP), is ubiquitous and
is difficult to avoid. Users of SOUP face uncertainties about the
quality of the SOUP or if the SOUP was developed with specific
properties in mind. To address these uncertainties, users often wish to
take proactive measures to establish the desired properties in SOUP.
Desired properties are customizations to SOUP as required by the user
based on particular concerns, policies, restrictions, emerging risks, etc.
The approach for establishing properties investigated in this research is modification of the SOUP itself. Without the program source code, these modifications have inherent risks. Specifically, SOUP modifications might break expected program semantics, not meet overhead constraints, or not properly establish the desired property. Users must balance these risks with the engineering costs and potential benefits that SOUP modification provides. These trade-offs and implementation decisions make SOUP modification a difficult engineering problem, and at present this problem has been minimally explored.
The proposed research investigates SOUP modification as an engineering problem and introduces a process for achieving SOUP modification, called speculation. This research will examine the guiding principles and practices necessary to produce a general engineering framework of speculation.