CS 525V (S02) : Introduction to Computer-Aided Verification

Exercises will be posted here as they become available. I expect to give roughly 4-5 exercises (in addition to the project) over the course of the semester.

General Assignment Documentation

• Notes on preparing SMV assignments

Assignment Handouts

• Assignment 1: Warming Up with CTL and SMV.
  Due: Jan 22, 2002, by class. Submit the two SMV files via turnin using assignment name smv-warmup. Also bring a hardcopy of the entire assignment to class.

  Auxiliary files:

  • simplemem.smv

• Assignment 2: Modeling Environments and Protocols in SMV
  Due: Feb 12, 2001, in class (hard copy) and electronically. Submit electronic versions of all files via turnin using assignment name hwk2.

  Auxiliary files:

  • itlc.smv

• Assignment 3: Modeling Simple Software in SMV
  Due:Thursday, March 14, 2002, 4pm via hard copy to my office or mailbox and electronically. Submit electronic versions of all files via turnin using assignment name hwk3.

  The 24 hour late policy does not apply to this assignment. All files/hard copies must be turned into me by 4pm on Thursday. If I don't have your assignment hard-copy in hand by 4pm Thursday, I won't grade it.

• Assignment 4: Modeling with Alloy
  Due:Tuesday, April 2, 2002 in class via hard copy and electronic submission. Submit electronic versions of all files via turnin using assignment name hwk4.

  The 24 hour late policy does not apply to this assignment. All files/hard copies must be turned into me by class on the due date. If I don't have your assignment hard-copy in hand at that time, I won't grade it.

• Assignment 5 (optional): Bounded Model Checking
  Due:Tuesday, April 23, 2002 in class via hard copy and electronic submission. Submit electronic versions of all files via turnin using assignment name hwk5.

  Due to time constraints on when I must submit final grades, I will not accept late assignments.

  This assignment is purely optional. The grade on this assignment will roughly replace a poor grade on an earlier homework.

Your final project will take one of the following three forms (your choice):

1. Modeling and verification of a substantial design using one of the tools that we discuss in class.

2. Modeling and verification of a reasonable design using a tool that we do not discuss in class. Your project should contrast your experience with the new tool with what you would have expected from verifying the same design in the tools that we used in class.

3. A survey paper (roughly 10-15 pages) describing and contrasting the state-of-the-art approaches to verification in a particular problem domain.

In general, these will be individual projects. However, I will consider team projects provided you propose a project of sufficient scale to warrant multiple team members.

In each case, you will turn in a report describing your project and give a brief presentation on your project to the class at the end of the semester (presentation duration to be determined once I know how many students are enrolled).

You must submit a short project proposal and have me approve it. Proposals will be due sometime mid-semester. Start thinking about the kind of project you would like to do. Possible domains include aspect-oriented verification, security protocols, e-commerce protocols, graphical-user interfaces, wireless network protocols, particular computer architectures, embedded software systems, etc. I'm glad to help you develop project ideas if you ask.

Proposal Details

By class on Tuesday, March 12, you must turn in a one-page project proposal.

• If you are doing a verification project in SMV, describe the problem you are going to verify, where you are getting the specification from, and what properties you expect to verify.

• If you are doing a verification project in a different tool, describe the nature of the tool and its special features (ie, what's different from SMV that makes it interesting to try), the problem you are going to verify, where you are getting the specification from, and what properties you expect to verify.

• If you are doing a survey paper, describe the topic you are surveying, provide a list of the papers you plan to read, and provide a list of the questions that you intend to address in writing your survey (ie, how are comparing these papers? -- these make your project a survey comparison rather than a book report).

Final Report Details

Final reports are due in hard copy to my office or mailbox by 5pm on April 26th (the last day of grad classes).

If you are doing a verification project (either in SMV/Alloy or in some other tool), your report should contain (not necessarily in this order):

• A description of the problem you are verifying.
• The desirable properties of that problem and which ones you are verifying in this project (any why you chose those).
• A discussion of how you developed the model for your problem (what modules, what interfaces, what does the environment do, etc). Include a discussion of what features of your design you chose to not model: why did you leave those features out?
• A discussion of what was easy/hard to model in your chosen tool
• If you used any advanced techniques (abstractions in SMV, for example), explain how and why you used them.
• If you are using a model checker other than SMV, explain how SMV would have made your project easier to harder to do (ie, contrast SMV to your tool in the context of your problem).
• A concrete summary of the verification status of each property.
• Include printouts of all models developed (and document the models).

If you are doing a survey paper, your report should contain

• An overview of the problem you are studying, where it fits into what we've discussed in class, and why it's important in verification.
• Concrete examples to illustrate the differences between the approaches you are contrasting.
• An assessment of the techniques you've compared, on both technical and practical grounds: e.g.: when would you use (or not use) the various techniques, what level of expertise is required to use the techniques, how do they fit into existing verification frameworks, etc.

In both kinds of projects:

• Quality of writing counts. Make sure your writing is clear and precise. Avoid vague statements.
• Proofread your report for spelling errors, as well as grammar errors (to the best of your ability based on your English writing skills).

Presentation Details

Project presentations will occur over the last two meetings of the course. Prepare a 15 minute presentation. We'll also have a few minutes for questions. Presentations should use transparencies or Powerpoint (or something similar).

• Week 1: Paul, Ji, Raj, Beth, Archie, David, Yakov
• Week 2: Mitesh, Amit, Larysa, Nikhil, Venky, Gleb, Dev, Sudeep

NOTE: If you use Powerpoint, we can run the talk off of my laptop. However, my laptop doesn't have a floppy drive, so you need to send me your presentation via email or a URL by 4pm on your presentation day.

When grading your presentations, I'll be looking for clarity, technical accuracy, and coherence. For example:

• How well you described the problem you are working on
• If you are using a new tool, how well you put the new tool in context with the ones we've used in class
• If you are doing a survey paper, how well you explain some of the different approaches you are studying
• Whether you used examples well in your talk (use lots of examples to make your talk concrete)

You do not need to explain every aspect of your project in the 15 minutes. A good presentation on part of the project will be better than a rushed, shallow presentation on the whole project. If you present only part of your project, make sure you present some of your own work (ie, if you are using a new tool, don't just give a tool demo -- also talk about the problem you are modeling and your experience with the tool on your problem).