PhD Program in Software Engineering Institute for Software Research, International

The ISRI offers a PhD Program in Software Engineering within the School of Computer Science.

Introduction

Software-based systems have established themselves as essential parts of business and everyday living in the emerging global economy. As software becomes ubiquitous, the relation between end users and software development undergoes fundamental changes.  No longer is software produced from scratch by a team of experts and delivered to clients. Increasingly,

• The software is developed by adapting existing components and services;
• The delivered systems are highly distributed and software-intensive but not purely software;
• The systems must be designed and fielded under economic and legal constraints;
• Clients are intimately involved in the development and configuration of the system; and
• The requirements for the systems emerge as the clients understand better both the technology and the opportunities in their own settings.

Carnegie Mellon has established the Institute for Software Research International (ISRI) as a center of excellence for long-term interdisciplinary fundamental research, apprenticeship-based education and training, and international collaborations to address the challenge of designing, developing, integrating, validating, and maintaining practical, large-scale, high-quality software-intensive systems. The creation of the ISRI is a natural extension of Carnegie Mellon's long-standing commitment to research in software systems.

ISRI research activities include experimental prototyping, empirical modeling, codification of experience, formal analysis, creating design/development strategies for modern software, and developing public policy positions.  The educational activities are tightly integrated with the research and demonstration projects.

ISRI's PhD program infuses the research agenda with deep appreciation of the practical issues of developing product-quality software-intensive systems.

Expectations for Graduates

Graduates will be prepared for faculty positions in software engineering, for research positions in industrial laboratories, and for leadership positions in development in both computer industry and application shops.

• As faculty, they will be distinguished by their genuine understanding of software design and development issues, which can be expected to shape their selection of research problems and evaluation of research results.
• As industrial developers, they will understand not only academic research issues but also engineering constraints that arise from public policy, economic, regulatory, and market issues.
• As senior software system developers, they will have a distinctive perspective that addresses specific problems in the context of the principles and results of the field.

In addition to talent and research skills that will allow them to advance the field, graduates will have a clue about practical issues of software design and development from requirements acquisition through maintenance. They will be prepared to enter research or advanced development positions in application areas, distributed networks, embedded/critical systems, and other specialties as well as conventional software system development. They will know enough about managing projects and about the software industry that they will not be naive about business questions.

[back to top]

Prerequisites

Evidence of proficiency in computer science at an undergraduate level with emphasis on development of software or hardware systems.

Evidence of intellectual ability to succeed in a rigorous, high-quality doctoral program, demonstrated through transcripts and GREs.

Prior industrial software development experience, especially software design and programming experience as a member of a system team, is a strong asset. Students without sufficient quantity or quality of experience will spend more time during the program gaining practical experience. At a minimum, we expect that every student entering the program will have experience equivalent to two summer internships as a member of an industrial software development team.

Extensive, high-quality software development experience may substitute for any of the normal requirements. It will be individually evaluated in each case.

[back to top]

Applications

Please make applications through the usual SCS admissions process. The application deadline, mailing address, and fee are the same as for other SCS PhD programs. In addition to the materials required for all programs,

1. Include a cover message asking to be considered in the Software Engineering PhD program. Also make a notation on the application form indicating that you are applying for this program.
2. Elaborate your professional resume with a description of how your industrial or commercial experience prepares you for this PhD program. For example, outline the parts of the software life cycle with which you have experience or exposure.
3. The GRE general test is required, and the computer science subject test is strongly recommended but not required.
   
   [back to top]

Program Plan

Estimated time requirement: 4 CY, accounted for as 1 AY for courses, 0.5 AY for teaching, 1 CY for background work and proposal, 2 CY for thesis, including contributions to sponsoring project. Expect up to 1 additional CY for students lacking sufficient prior experience.

A note on highly experienced students: We especially value the special perspective that very senior software developers can bring to a research program. These very senior developers typically have 10 years or more of system development experience, at least 5 years in a position of senior responsibility for design and outcomes. It is rare to attract such a person as a PhD student. Instead of designing a program that anticipates all such situations, we recognize that each will be a special case requiring special arrangements.

[back to top]

A Typical Plan of Study

Evaluations of student progress in SCS at Carnegie Mellon compare student achievement to the criterion, "Is this student making reasonable progress to completing an excellent PhD?". Consequently, the requirements may be satisfied in whatever order best serves the student and the institution.

A typical plan of study includes the following requirements:

• A course requires about 12 hours/week
• Participation in the weekly ISRI Software Research Seminar requires about 3 hours/week plus one presentation preparation
• TAing is about a half-time load
• A student should usually be devoting about half time to a research project. This leads to a thesis proposal in spring semester of the third year and completion of the thesis in the fourth or fifth year.

Common variations include TAing a course in the Spring of year 1 or 2 instead of the Fall of year 2, TAing a second course instead of mentoring a studio for a year, and moving the course from the TAing semester to the third year.

[back to top]

Courses

The courses listed below are examples of courses available. Students interested in courses not listed here should contact their advisor for guidance and suggestions.

MSE Core Courses: We have adapted some of the existing Master of Software Engineering (MSE) core courses to serve both MSE and PhD students. MSE courses available for PhD credit are cross-listed as 17-7xx (e.g., 17-751, Models of Software Systems), and may require an additional project to satisfy the PhD requirement.

Design and Engineering

• 17-752 Methods of Software Development (Adapted from MSE Core Course 17-652)
• 17-755 Architectures for Software Systems (Adapted from MSE Core Course 17-655)
• 17-939 What Makes Good Research in Software Engineering?

Systems

• 15-712 Advanced Operating Systems and Distributed Systems
• 15-740 Computer Architecture
• 15-744 Computer Networks
• 18-749 Dependable Embedded Systems
• Application systems courses include systems courses offered through the Language Technologies Institute (LTI) in the School of Computer Science

Analysis

• 15-750 Algorithms Core
• 17-751 Models of Software Systems (Adapted from MSE Core Course 17-651)
• 15-812 Semantics of Programming Languages
• 15-814 Type Systems for Programming Languages
• 15-853 Algorithms in the Real World

Business and Policy

• 17-910 Business Models for Software Development Methods
• 90-802 Information Security: Comparison of US and European Policies

Electives

• 15-887 AI Planning, Executing and Learning
• 17-810 Empirical Methods in Software Engineering Research (6 unit mini-course)
• 17-811 Self-Healing Systems (6 unit mini-course)
• 17-812 Open Source Software Development (new 6 unit course)
• 17-993 (listed internally as 17-960): How to Write a Good Research Paper (6 unit mini-course)
  
  [back to top]

Practical Experience within the Program

An integral part of the ISRI research program is ongoing interaction with industrial-strength software development in a real (not just realistic) setting for hands-on education.

Two kinds of practical experience are required: 12-unit practicum and experience in lieu of prior industrial experience.

Examples of a 12-unit practicum include:

• Longitudinal: short project from design to implementation, including course-based examples (15-822) or projects arranged with industrial partners.
• Topical:short exposure to aspects of the topic on several different projects. For example, a quality control practicum might spend 3 weeks each with projects doing design reviews, test plans, user studies, and change requests.
• Specialty: participation in the aspect of one or two projects that's related to the student's chosen specialty. For example, a student interested in real-time systems would participate in the real-time requirements analysis, design, and validation of two or three projects.

Experience in lieu of prior industrial experience:

In addition to the Practicum, students who have had minimal prior industrial experience will be expected to deepen their understanding and perspective of industrial software development during the Ph.D. program. Alternatives include part-time work with local industry, extended internships in development groups (summer plus semester), and special arrangements with industrial partners in ISRI research projects.

[back to top]

Research and Thesis

Research in the program will address topics of interest to ISRI. Our focus is on systems that exploit the growing infrastructure for high performance, nearly ubiquitous computing and communication, especially systems that the public depends on for services provided through the electronic marketplace. The research approach for each project will be selected to match the needs of the project. Approaches appropriate to PhD theses include (but are not limited to):

• Novel methods for software development
• Implementation techniques for novel applications
• Automated support for software activities
• Measurement techniques for system evaluation
• Descriptive models that generalized from practical examples
• Guidance for making classes of design decisions
• Empirical models with predictive power
• Analytic models that permit quantitative or symbolic analysis
  
  [back to top]

For more information about the Software Engineering PhD Program, contact: isri-phd@cs.cmu.edu.

This page is part of the Institute for Software Research, International site in the School of Computer Science at Carnegie Mellon University. Please send inquiries about the Institute for Software Research, International to the ISRI office. This page was last modified on September 17, 2003.