What makes Useful Research (for Software Engineering PhDs)?

What Makes

Useful Research?

Your PhD Thesis

(PhD version)

Andreas Zeller, Saarland University

Artwork by Jorge Cham, www.phdcomics.com

The

Conception

The

Assessment

some cool idea

that might rock the world

that would put a dent in the universe

The Universe

Was it hard?

Is it elegant?

Is it useful?

or: What can be expected from you?

or: Who will be impacted?

or: What will its impact be?

Does the problem actually exist?

What do you improve?

Productivity? Quality? Risk?

Does it solve the problem

in principle or in reality?

At which cost?

From The Illustrated Guide to a Ph.D., by Matt Might

How do you measure this?

What assumptions does your solution make?

How do your results generalize?

How much change will your solution require?

What are the alternatives?

abstract

principle

ad-hoc

solution

Take risks

few

assumptions

lots of

assumptions

maximum distance

​​

general

revolution

incremental

reform

You

short term

long term

Thesis

Everyone

you?

Proving Android Apps*

Automatically

Predicting Bugs

from File Size

Code Completion

for Aspects

Flawed Reviews

D. Auteur

Université XXII

Paris-la-Nouvelle, FR

A. and N. Author

Big Ring Labs

Endless Loop, CC

Univ.-Prof. Dr. E. Autor

Universität Irgendwo

Neustadt, DE

Overloaded reviewers favor

Abstract. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Maecenas convallis nec ipsum nec aliquam. Mauris lacinia lacinia libero. Mauris vel vestibulum nisl, viverra porttitor lorem. In hac habitasse platea dictumst. Ut ullamcorper, augue sit amet eleifend blandit, elit tortor egestas purus, et luctus ipsum diam nec nibh. Sed ut eleifend ante. Aenean ullamcorper diam sed viverra pellentesque. Quisque laoreet vehicula ipsum, in consectetur felis suscipit gravida. Sed interdum sollicitudin ultrices. Curabitur tempus a lacus vitae aliquet. Nulla mauris purus, placerat eu nisl non, porta congue magna. Etiam ut risus sapien. Donec blandit mauris sit amet venenatis pulvinar. Sed viverra turpis id tempus pellentesque. Etiam non tincidunt ipsum. Phasellus lobortis, lacus sit amet fringilla hendrerit, massa tortor euismod velit, nec iaculis quam tellus in est.

1. Introduction

Phasellus tempus, felis vitae porttitor facilisis, felis diam tristique urna, in tincidunt leo ligula sit amet risus. Aenean ac eros in arcu aliquet iaculis. Curabitur gravida varius porta. Etiam adipiscing nulla sit amet scelerisque consequat. Integer id mi interdum est euismod pellentesque. Suspendisse laoreet in nulla in scelerisque. Nulla porta pellentesque risus eget faucibus. Vestibulum posuere purus ipsum, ac viverra lorem gravida eu. Quisque feugiat sed justo vitae tincidunt. Aliquam commodo, tellus ac vulputate auctor, nisi massa accumsan urna, nec sagittis augue dui a libero. Sed placerat elit gravida tellus porttitor, eget sollicitudin mauris dignissim. Nullam volutpat, mauris quis rhoncus accumsan, risus massa tempus enim, et facilisis diam risus id odio. Donec iaculis risus a sem eleifend, eget venenatis diam ultricies. Donec aliquet ut sem ut malesuada. Phasellus vel vestibulum neque.

Abstract. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Maecenas convallis nec ipsum nec aliquam. Mauris lacinia lacinia libero. Mauris vel vestibulum nisl, viverra porttitor lorem. In hac habitasse platea dictumst. Ut ullamcorper, augue sit amet eleifend blandit, elit tortor egestas purus, et luctus ipsum diam nec nibh. Sed ut eleifend ante. Aenean ullamcorper diam sed viverra pellentesque. Quisque laoreet vehicula ipsum, in consectetur felis suscipit gravida. Sed interdum sollicitudin ultrices. Curabitur tempus a lacus vitae aliquet. Nulla mauris purus, placerat eu nisl non, porta congue magna. Etiam ut risus sapien. Donec blandit mauris sit amet venenatis pulvinar. Sed viverra turpis id tempus pellentesque. Etiam non tincidunt ipsum. Phasellus lobortis, lacus sit amet fringilla hendrerit, massa tortor euismod velit, nec iaculis quam tellus in est.

1. Introduction

Phasellus tempus, felis vitae porttitor facilisis, felis diam tristique urna, in tincidunt leo ligula sit amet risus. Aenean ac eros in arcu aliquet iaculis. Curabitur gravida varius porta. Etiam adipiscing nulla sit amet scelerisque consequat. Integer id mi interdum est euismod pellentesque. Suspendisse laoreet in nulla in scelerisque. Nulla porta pellentesque risus eget faucibus. Vestibulum posuere purus ipsum, ac viverra lorem gravida eu. Quisque feugiat sed justo vitae tincidunt. Aliquam commodo, tellus ac vulputate auctor, nisi massa accumsan urna, nec sagittis augue dui a libero. Sed placerat elit gravida tellus porttitor, eget sollicitudin mauris dignissim. Nullam volutpat, mauris quis rhoncus accumsan, risus massa tempus enim, et facilisis diam risus id odio. Donec iaculis risus a sem eleifend, eget venenatis diam ultricies. Donec aliquet ut sem ut malesuada. Phasellus vel vestibulum neque.

*without pointers, parallelism, loops, and recursion; see future work

• papers in known contexts
• incremental papers
• easy to read papers

Was it hard?

Is it elegant?

Is it useful?

or: What can be expected from you?

or: Who will be impacted?

or: What will its impact be?

Bean Counting

Does the problem actually exist?

What do you improve?

Productivity? Quality? Risk?

Does it solve the problem in principle or in reality?

At which cost?

How do you measure this?

What assumptions does your solution make?

How do your results generalize?

How much change will your solution require?

What are the alternatives?

abstract

principle

ad-hoc

solution

lots of

assumptions

few

assumptions

​​

general

revolution

incremental

reform

You

Administration likes metrics:

short term

long term

• easy to measure
• easy to compare
• require no expertise at all

But I published something!