[Colloq] Thesis Proposal by Pengcheng Wu, Tuesday 7/26, 10am

[Rachel Kalweit]

College of Computer and Information Science

PhD Thesis Proposal Presentation by:
Pengcheng Wu

who will speak on:
Efficient and Expressive Aspect-Oriented Programming Languages

Tuesday, July 26, 2005
10:00am
366 West Village H

Abstract
Pointcut designators (PCDs) are key linguistic elements of AspectJ-like 
Aspect-Oriented Programming Languages (AOPLs). Compilation of PCDs has 
attracted significant research efforts in the AOPL community. The first 
part of this thesis work will address issues of compiling the CFLOW 
pointcut designator, which usually brings the most significant run time 
overhead among all of the PCDs. The current CFLOW compilation approach 
generates inefficient target code for multithreading programs and 
delivers unexpected behaviors under some circumstances, among other 
problems. Therefore, I propose a new compilation approach for the CFLOW 
pointcut, called cFlow-Passing-Style (FPS) to improve the current 
practice. With this new approach, for CFLOW intensive programs, we can 
expect better run time efficiency, correct behaviors according to its 
semantics and more broadly deployable target code.

The second part of this thesis work proposes the notion of Shadow 
Programming. In AOPLs, shadows are abstractions of static properties of 
run time execution events called join points. So far, shadows have only 
served the purpose of building AOPL compilers, and they are not 
accessible to programmers. The notion of Shadow Programming argues that 
shadows should also be made accessible to programmers at compile time so 
that those abstractions can also be exploited for a lot more tasks. To 
show the usefulness and feasibility of this notion, two compile time 
facilities, called Statically Executable Advice and Pointcut Evaluator 
respectively, are developed to use the exposed shadow information to 
support customized compile time analysis and to support more 
sophisticated join point selection mechanisms.

Thesis proposal committee:
Karl Lieberherr, advisor
Gene Cooperman
Mitch Wand
Shriram Krishnamurthi, Brown University