CS595 Applications in PL Final Project

General Info

• The project may be done in groups of 1-3 students. Expectations for the scope of the project will be scaled according to group size, so smaller groups aren't at a disadvantage.
• The goal of the project is to apply the tools and techniques learned in the class to a domain of interest to you, whether that's your research, another interest in CS or an interest outside of CS.

Project Scope

(Added 11/1) At this point, your group has discussed project ideas with Stefan at least once and you should have some general ideas of what project-specific results to include. As a broad guideline, each project should include:

• A definition of a model for some language, system or idea (syntax defining the language for your model, as well as rules for one or more judgments, which might include typing rules, operational semantics, cost semantics and others).
• At least one proof, roughly on the scale of the progress/preservation you've been doing for homework, that shows that your model "makes sense". For a language, this might be a proof of type safety (progress and preservation). Or it may be something else.

Project Ideas

1. Model the syntax and semantics of an existing language or system and develop a type system to rule out some class of "bad" programs (Examples: the division by zero type system from HW1)
2. Develop a syntax and semantics to model something that isn't a (conventional) programming language, and prove something interesting about it (this could involve developing a type system as in Idea 1, but doesn't need to). (Examples: the model of red-black trees and corresponding proof from HW0, this paper)
3. Extend a language model from a paper (add features, rule out more "bad" programs, allow more "good" programs, etc.)

If you have other ideas that you think would make a good project in line with the goals of the class, talk to Stefan.

If you're looking for inspiration or aren't sure what fits within this, any of the papers linked under the "Applications" section of the course site would make an A+ final project (keep in mind, of course, that most of these are published conference papers, so are more thorough than would be expected for a final project.)

Timeline/Deadlines

These are highly subject to change.

Late days may be used on the proposal and final paper. Note that for group projects, late days will be charged to everyone in the group with late days remaining (that is, you can still use late days as long as anyone in the group has late days remaining, but you can't pool them.)

Deliverables

Initial Idea

1. Find a group for the final project (or decide to work alone).
2. Schedule a meeting (for your whole group) with Stefan to discuss the idea.
3. Submit 2-3 paragraphs (as a group) describing the idea for your project.

The submission doesn't need to be very formal but should indicate some thought about what you plan to do and how you plan to relate it to the course topic. For example, if you're solving a problem in a language or system, describe the language or system and explain the problem. If you're modeling something outside of CS, talk about the thing, what features of it you plan to model, and some ideas for what you plan to say/prove about the model. If you're planning to extend a paper, cite and summarize the paper and talk briefly about your idea for an extension.

The submission won't be graded; you get full credit by submitting something reasonable by the deadline.

Proposal

Your proposal should be approximately 1-3 pages and formatted like a research proposal with at least the following general outline.

1. Introduction
   • Motivate the problem your project aims to solve; why is it important?
   • Give an overview of your proposed solution.
2. Preliminary Work
   • Describe any work you've already done toward building your model: if you have any (rough draft) syntax or rules, present it here.
   • The focus should be on the text description of what you've done and how it achieves your goals, not the low-level technical details of how you've accomplished it (intuition behind your syntax/rules and examples are great!)
3. Proposed Work
   • Outline what is left to be done.
   • Give rough statements of theorems/lemmas you plan to prove. Perhaps describe at a high level how you might go about proving them.
4. Timeline and Research Plan
   • Give a (roughly) week-by-week timeline of when you plan to accomplish each of the goals of the proposed work.
   • For group projects, describe who is going to do what.
   • Example: "(Week of Nov. 15) Group members A and B are going to work together to flesh out the operational semantics. (Week of Nov. 22) B will prove Theorem 1. (Week of Nov. 29) A and B will make slides for the class presentation (Dec. 2-Dec. 6) A will prove Theorem 2, B will begin writing up the other results. (Dec. 6-Dec. 8) A will write up Theorem 2, A and B will polish the writing of hte paper."

Final Paper

Your final paper should be in the format of an ACM research paper, with some modifications detailed below. This guarantees that everyone's paper will be in the same form and page limits, etc., will be fair. ACM provides their own LaTeX class file (as well as a Word template if you prefer to use that). Information about the ACM class file and Word template is on the ACM's website. If you use Overleaf, this page has details about how to use the ACM class on Overleaf. If you're using LaTeX (locally or on Overleaf), I put together a template that has everything you should need.

I'll discuss the format and how to write a research paper in more detail in class, but here are some highlights and differences from a standard research paper.

You should have:

• An introduction, much like the intro of your proposal.
• One or more sections that discuss your contribution in a readable, user-friendly way:
  • Give all of the formal grammars, definitions, judgments, rules, proofs, etc., needed to know what you did, but for each block of formal mathematical notation, also explain what it means/does in plain English.
  • Give examples to help illustrate how your model works, how to use it, etc.
• A conclusion.

In contrast to most standard published research papers, you don't need:

• An abstract
• A discussion of related papers

Your paper should be a maximum (not minimum) of 10 pages in the ACM format. If you really need more space, let me know why and I may make an exception. There's no minimum page limit, but make sure you fully describe your project and give all of the examples, intuitions, etc., you feel are needed.

For Writing Help

The IIT Writing Center (see blurb below) is a useful resource if you're looking for more guidance or feedback on your proposal or final paper. The tutors there should be able to help you. Full disclosure: One of them (Hannah) is Stefan's fiancee.

Consider using the campus writing center as a resource for help with written assignments like research papers, technical reports, or presentations. The writing center provides one-on-one conferences for both undergraduate and graduate students. Tutors can assist with any stage in the writing process, from brainstorming and outlining to final edits and bibliographies. Tutors read academic writing with a critical eye given not only to grammar and syntax, but also to understanding assignments, paragraph structure, clarity of ideas, strategies for revision, and even strengthening arguments. Visit here to make a (free!) appointment: https://www.iit.edu/humanities/student-resources/writing-center.

Grading

Final project grades will be based on the following:

• Initial idea (full credit if submitted on time).
• Proposal
• Applicability to course
• Appropriate scope for size of group
• Correctness of proofs
• Communication of ideas (in class and the final paper)
• (For group projects) Did group members contribute equally?

Rubric for final presentation and project