Instructor: Prof. Seth Fogarty
Email: sfogarty@trinity.edu
Office: CSI 270M
Meets: TuTh 12:45-1:45, CSI 356
Salon - noun (sa·lon) /ˈsælɒn/ /səˈlän/:
(in the past) A regular meeting of writers, artists and other guests. Often following Horace's
definition of the aims of poetry - aut delectare aut prodesse, “to please or to educate.”
Informatics - noun (in·for·mat·ics) /in-fər-ˈma-tiks/:
The study of computational systems.
In this salon we will read a selection of computer science papers with the theme of "surprising results." Papers will range from "Reflections on Trusting Trust" by Ken Thompson in 1984, through "Termination Analysis for the Untyped Lambda-Calculus" by Neil Jones and Nina Bohr in 2004, to "ChatGPT is Bullshit" by Michael Hicks and Jason Humphries in 2024.
In addition to understanding the breadth of award-winning research in computer science, students will learn how the academic study of computer science is conducted. By the end of this course you will be able to:
Skim a paper to determine if it is relevant to a topic or question you are researching.
Articulate the high-level results of a paper, even if you do not have the background to understand the details.
Identify which claims or results in a paper are from existing results, which are novel, and what future work is afforded by the results.
Read an academic paper in Computer Science in depth.
Find related works to a paper, both to understand the background of the paper and to see what work has followed it.
Write a useful summary of a paper, so you can find it again if you need it.
This course meets twice a week. Class time will be spent reading papers in the round, with frequent stops to clarify or discuss content, find resources on unfamiliar elements, and connect the paper to the computer science curriculum. While the majority of time for this course will be spent reading in class, we will collaboratively write a document summarizing the papers read during the semester, and each participant will delve into a topic of interest.
Each participant will take one paper discussed during class to summarize in a living bibliography. These summaries should include not just the bibliographic information, but also significant results, important theorems or constructions, and other useful elements that a researcher might need to cite in the future.
Finally, each participant will select one topic of interest to delve into and produce an artifact about. A great deal of flexibility is given on the topic selected, the mode of research, and the deliverable produced. The following is the suggested process - if your mode of research will be significantly different, please check in with Dr. Fogarty.
Selecting a Topic: your topic could be directly from a paper read in class, something mentioned in a paper, the work of a Turing Award winner not covered, or a topic of interest from another class.
Finding Related Work: you should make a list of roughly 10 papers whose title or citations suggest they are applicable to your topic. You should not yet read these papers in any depth, but may look at their citations for other papers.
Identifying Key Papers: by reading the abstracts of the papers, identify 2-4 that seem particularly useful. These papers may not be the foundational papers, or even papers with new results. Survey papers are likely ideal for your purposes.
Skim and Summarize: skim the ‘key papers’ identified above and summarize their contents. The goal of this step is not to understand the papers in depth, or even be able to articulate exactly what they cover, but to further narrow your focus.
Read in Depth: identify which paper(s) or sections of papers have the most important information on your topic, and read them in depth. This should include between 10 and 40 pages of in depth reading.
Produce an Artifact: you should prepare an artifact suitable for an undergraduate audience, equivalent in scope to a 400-600 word essay. This may be a written document explaining one of the algorithms or constructions with examples, a short video motivating the value of an approach, an argumentative essay about work that was overlooked or undervalued, or another deliverable of your choice.
In addition to the artifact, participants will give an informal 5-10 minute presentation on your topic during the final exam period (Wednesday, May 7th from 12:00-3:00). The aim of this presentation is to interest your audience in the topic - not to explain 10-40 pages of densely written academic writing in five minutes. Visual aids are optional, but if included should be appropriate to the goal of the talk.
The course grade will be based on four categories, given the following weights:
Participation - 60%
Paper summary: 10%
Delve: 25%
Informal Presentation: 5%
The grade for the paper summary and delve are primarily derived from self-assessment and reflection, which will be completed towards the end of the semester. The presentation will be graded on a simple rubric, which will be made available ahead of time.