Project

The project is composed of three parts with a different deadline:

• the repo creation (creating a github repo on the classroom)
• the artifact (code you produced)
• a report
• the review of 2 artifacts produced by someone else

You must do all four parts to pass the course!

Dates

Hint: the “repo creation” is only there so that I have more time to assign reviewers as I will at a conference the week after.

Themes

We are completely opened on what you write about. The only constraint is that you should go in depth on two themes from the lecture.

Some ideas:

1. Develop a plugin for one of the text editors we discussed.
2. Evaluate several plugins based on your specific use case and explain their differences and inner workings.
3. Replicate a favorite book’s style in LaTeX.
4. Automate analysis using Bash or Python.
5. Implement automated testing for your side project, configure Git with hooks, and compare it to Jenkins.
6. Conduct extensive LLM testing to assess differences.
7. Create complex yet readable plots with gnuplot/jupyter. Make sure to include some of things you tried in the report.

Generally, just using the default option of a tool is not sufficient to claim having used a tool!

To Submit

As an example, we have written an example project.

Repo creation

This part is simply accepting the assignement from the Github classroom (link will be on the Ilias forum). This is useful for us to decide how many people submit (and for me to assign reviewers as I am at a conference with a bit more time).

We will not look at the current status of the repo before the artifact deadline.

Report

The report should be:

• Approximately 8 to 12 pages long, excluding references and the table provided in the example project;
• Include a table at the conclusion that showcases the sources utilized from the lecture (refer to the example project);
• Be written in either English or German;
• Address two topics discussed in the lecture.

We do not want to see a report written by your favorite LLM that you did not bother reading. LLMs tends to write very general sentence with extremely little information and with no relevance to your report.

Artifact

As part of the report you will produce code / bash scripts. You should produce an artifact with a proper README that runs under the following VM https://zenodo.org/records/7113223.

If you need extra proprietary software or Windows-specific software (Mathematica / Maple / …), please tell us early.

The artifact should include:

• a draft version of your paper;
• a license;
• include a proper README that shows what you have done and what part of the paper you are trying to produce. You are allowed to ask for installing extra software via the standard apt get;
• be runnable in less than 10 minutes. If it takes more than that, you should provide a small (meaningfull) subset of your input.
• you should not include the OVA image, only a folder
• the folder will be imported in the image (no shared folder)
• include the ouput your tool produces to be able to compare

Unlike the instructions on https://zenodo.org/records/7113223, you are allowed to use apt-get, but you should assume that you start with an empty virtual machine.

Review

You will be assigned two artifacts by someone else. We will assign them on the 16th or 17th. You should write a small review including:

• a few sentence what the artifact is about
• whether you managed to reproduce what is done in the artifact and if that makes sense w.r.t. the results of the draft.
• some comments on the code

Some more information:

• do not expect to get exactly the same timings, try a few to find how much faster/slower hardware is and see if you stay in the same ballpark.

• we will use GitHub. Put the report in an issue.

We expect around half a page of review. However, you are not expected to debug or guess what the artifact is about.

You do not have to use the virtual machine… but if there are any problems, you need to check the virtual machine.

Changes

• There is now an extra deadline to create the repository on the Github classroom

• We require 8 to 12 pages (instead of 6 to 8)

• If you got graded, you must change the artifact and the report entirely. We do not want to see the same topic again.

• If you did not get graded (because you were not registered to the exam), you can reuse what you did. However, do not forgot to adapt the draft of your report to the new length.