Description
This talk will describe efforts in finding a unified framework that generalizes several conceptually similar program analyses, e.g. all known variants of size-change termination, and other analyses based on interpreting computation path-effects in finite abstract domains.
Is it possible to isolate shared underlying principles to these analyses? The current work aims to identify common algebraic structure of which these several essentially similar analyses arise as special cases. It is the author's hope that this will be beneficial in the following:
A common domain may allow us to compare the analyses in a rigid fashion.A common basis for the analyses can function as a template for constructing new program analyses that build on abstract summaries of computation path effects.Because the arising algebraic structures are well-known, we may be able to apply classical algebraic results to give insight into properties of the analyses.New results that apply to the general domain may be shared by all of the analyses.| Period | 21 Apr 2006 |
|---|---|
| Event title | The Second DIKU-IST Joint Workshop on Foundations of Software |
| Event type | Conference |
| Organiser | Graduate School of Information Science and Technology, University of Tokyo |
| Location | Shonan Village Center, Kanagawa, Japan, JapanShow on map |