Research and Development

" The gap between theory and practice in theory is nowhere near as big as the gap between theory and practice in practice. "
- Anonymous

Members of DOPL Research Group: Vicky Singh, Mehmet Kaya, Jim Fawcett, Phil Pratt-Szeliga, Mike Corely, Mubarek Mohammed, Jing Jia, and Kanat Bolazar

DOPL Lab - CST 4-116

Celebrating Kanat Bolazar's Completing his Ph.D. Program
Friend Liz, Kanat, Phil, Mubarek, and Me

ESF Modeling Group:
Jim Fawcett, Yu Chen, Shannon Connely, Yang Yang, Tom Taggart

Research advisees and Teaching Assistants at dinner

 

Early Members of the DOPL Research Group

I started a small research group that focuses on software complexity, accessibility and reuse. Some of this work centers around the analysis and evaluation of large software systems using grammar-based static dependency analyses (Research Statement). We are also interested in the notion of a Software Foundry and an associated Foundry Object Model, developing tools to support understanding large systems, and using distributed and massively parallel machinery to solve challenging engineering and scientific problems. The research group currently consists of four Doctoral students, several students preparing Master's Theses under my direction, and a number of students doing Independent Studies.

In the photos above, Kanat Bolazar is working on ideas and tools to investigate the behavior of large existing software systems, Phil Pratt-Szeliga is interested in distributed and parallel means to attack challenging computational problems, Mehmet Kaya is currently preparing for his qualifying examinations, and then expects to work on restructuring software, Mubarek Mohammed is working on software development visualization models, and John Goddard is just completing Master's Thesis research on Voice Activated Studio Recording.

Only students, known to me through work in one of my courses, are accepted into this group. Because of my teaching load I do not seek outside funding, although we occasionally work on funded projects that happen to come our way. I fund Doctoral and Master's Thesis students by making them Teaching Assistants in my courses. For the reasons, cited above, I do not reply to external requests for funded research under my direction.

Software Accessiblity

A major concern of ours is to honor the connectedness of software. That is, we provide accessiblity to code, its documentation, and test drivers, for any component, including any lower level components on which it may depend. The implications are that code, documentation, and test are distributed across a network of components and are reusable at any level of the distribution graph.

Some of this research extends into the classroom:

• Students in CSE681 - Software Modeling & Analysis have investigated collaboration infrastructures, e.g., a system composed of Clients and Collaboration, Repository, and Test Harness Servers.
• Students in CSE687 - Object Oriented Design have implemented test systems composed of Clients and a remote test harness.
• Students in CSE681 - Software Studio have collaborated as a team (each year) to build interesting systems: Virtual Repository Testbed, Collaboration Server, Remote Software Assistant, Repository Testbed, Project Center, Cloner, and Software Foundry.

Ph.D Dissertations:

At this time, one student from the DOPL group, Murat Gungor, has successfully defended doctoral research: Structural Models for Large Software Systems. Six other students are currently working on their doctoral research in the DOPL Group:

• At this time, one student from the DOPL group, Murat Gungor, has successfully defended doctoral research: Structural Models for Large Software Systems.
• Kanat Bolazar just completed doctoral research on means to explore and understand large software systems by dynamically posing contracts and, through mutation testing, verify their quality.
• Phil Pratt-Szeliga has completed and successfully defended a Master's Thesis on consuming large scientific programs on Graphic Processing Units (GPUs). The focus of his doctoral research is the mapping programs written in high-level languages like Java onto resident GPU hardware. The current state of the art requires a lot of knowledge about the machine architecture and GPU characteristics and a lot of low-level programming to consume a program, dividing the computations between host processor and GPU infrastructure. Phil's goal is to automate most of this so that a scientist has realistic hope of seeing significant performance improvements without a lot of background knowledge about the computer architecture.
• Mehmet Kaya's focus is on the restructuring of, possibly large, software systems by extracting classes and functions, inserting interfaces, and repackaging source code.
• Mubarek Mohammed has begun working on visualization techniques for disclosing and understanding the structure and operations of large sofware systems. He is working on ideas related to visualizing the operation of threads and exploring ways to disclose the static structure and dynamic dialogs between components of large systems.
• Vicky Singh, a Master's candidate, is working on an interesting project this summer in support of the work our group is doing, especially related to that of Mubarek in visualization. Vicky is developing a prototype of a program automation tool, designed to show the evolution of execution through an animated callgraph, showing the progress of threads through the structure in playback and single step mode.
• Jing Ja is working in Computational BioInformatics. She is now preparing for her QE2 examination, and will begin her own research following that.
• Mike Corely is interested in securing large software systems. He is now preparing for his QE1 qualifying examination and will begin scolarly research when that has been accomplished.

Master's Theses:

1. Debugging using Virtualization Hooks Ryan Wilson : Enhanced Debugging Through Virtualization. Completed December 2011.
2. Scientific Computing Using GPUs Phil Pratt-Szeliga : Automatically Utilizing Graphics Processing Units from Java Bytecode. Completed April 2010.
3. Software Restructuring, Santosh Singh Kesar : Automatic extraction of functions from large weakly structured code. Driven by rule-based static analysis. Completed October 8, 2008.
4. Model-Driven Development using the Software Matrix Technology, Tilak Patel : Model-Driven development generates code from UML style models. This becomes much more than a code generation wizard when there is a strong concept of components, the Software Matrix, used as an integral part of this process. Completed December 21, 2007.
5. Cross-Platform Development using the Software Matrix Technology, Vijay Appadurai : Cross-Platform Development targets flexible implementation on more than one platform. Vijay used Linux and Windows platforms for his demonstration. Completed March 23, 2007.
6. Self-Healing Systems using the Software Matrix Technology, Anirudha Krishna : a self-healing framework using the Software Matrix structure. It is effective and remarkably simple. Completed December 9, 2005.
7. Aspect Oriented Programming, Ramaswamy Krishnan-Chittur : explores several interesting applications supported by .Net interception. Completed April 19, 2005.
8. Software Matrix, Riddhiman Ghosh : an experiment using message-passing and broad use of the observer pattern to make software salvage a useful paradigm. By salvage, we mean the lifting of a significant block of one system to be used within another without terminal arterial bleeding. Completed December 15, 2004.
9. C++ Beans, An Experiment in Simplification, Siddhartha Azad, completed October 2002
10. Dynamic Internet Communities : A Comparison between Grid and Peer-to-Peer Computing, Shawn Dahlen [converted to an independent study - September 2003].
11. Performance Comparison of .Net and J2EE for some typical remoting and web service applications, Srinivas Shilagani [converted to an independent study - May 2004].

Development Projects