Amherst College
Department of Mathematics and Computer Science
Computer Science Thesis Proposals
Computer Science majors seeking to qualify for honors must
submit a thesis proposal during the spring semester of their junior year.
Proposals can be submitted to the department secretary, Anne Torrey.
The deadline for the proposal is 5:00 pm the second
Friday after Spring Recess. There will be no extensions to this
deadline (with rare exceptions of unusual and dire circumstances).
Proposals will be reviewed by the Computer Science Faculty: those
that demonstrate the desire and ability to
engage in a research and writing project on a significant
topic of Computer Science will be approved. Authors will be notified
about acceptance or rejection of their proposals (and about which professor
will supervise the research) before the week
of preregistration for fall courses. Authors of approved proposals
are permitted to enroll in Computer Science 77, the senior honors
course for the fall semester.
Requirements for Thesis Proposals
The proposal document should present a research question or topic that
might be pursued as a senior thesis. Topics may be suggested by
faculty members, found in journals and research publications, developed
as extensions of coursework, mined from past student theses,
or continued from summer employment.
Juniors who are considering honors work are strongly encouraged
to attend the Senior Honors Research Presentations held in
early March, and to visit Computer Science faculty members early in the
spring to discuss topic ideas, research expectations, and suitable formats
for thesis proposals. Students are also invited
to present draft proposals to faculty members for comment.
The proposal should be no more than 10 pages of double-spaced, 12
point type, not counting the bibliography. The document should
describe clearly the proposed project: specifically, what is the
problem to be addressed, and what is the context of the problem?
Some previous work should be surveyed to demonstrate that this
problem has been investigated by other researchers, and
that outstanding questions remain. The proposal should then address
which aspects of the problem could be approached in a new manner,
and outline at least one possible approach to studying the problem.
Alternatively, a student may propose to survey a topic, with
less emphasis on original research and more emphasis on review of
published articles. In this case, the topic should be presented together
with evidence of substantial background work to be investigated. The
author should explain the value of collecting and presenting this
background material in an expository format.
The proposal will be evaluated on four criteria:
- The topic: Is this a significant computer science topic?
Is the topic proposed of a reasonable scope for a senior thesis?
For example, ``The Whole New OS (WNOS),'' while a
potentially fascinating project, is far too broad for
a one-year investigation. ``A modified thread scheduling algorithm
for server workloads, examined in simulation'' is a more tractable topic.
- Background: Has the author found any existing published work on
this topic? What other approaches to this topic have been
taken? Is it a solved problem, or is there more to do?
While an exhaustive presentation of all related work
is not required, the proposal should provide evidence that the author
has read and understood one or two recent articles on the topic.
- A proposed strategy: What new approach is to be taken? Most
importantly, what will we learn from this new approach? In the
case of a proposal for original research, the author should explain
how this proposed approach may reveal something new about the subject.
In the case of a proposal for a survey thesis, the author should show
how the questions posed and answered by the thesis will contribute to
our understanding of the subject.
- Writing quality: Is the author able to present the above issues clearly
and succinctly? The proposal should show impeccable spelling and
grammar. The proposal should be well structured, with terms defined,
ideas clearly described, and past work accurately summarized.