Mystery Languages


The Mystery Language (ML) thread has you understand the design space of specific language features. In an ML assignment, you are given a new language feature’s syntax and a very loose description of it. The assignment will contain (say) three different languages that contain that new feature, each presented in terms of a black-box implementation of that language.

Each language will contain this new feature but will define it to behave in a somewhat different way. You will need to use the implementations to explore how the feature varies across the languages and try to pin down the differences precisely. You ultimately produce a classifier: a small set of programs that, collectively, tell the different implementations apart by producing different results under each implementation.


After two hours, come talk to the staff and show them what you’ve tried so far. Chances are your exploration attempts aren’t bearing fruit and you need redirection.

Grading Standard

You need to classify most of the languages on most of the assignments. (This has not been a problem for students in the past.) You might not get all of them, but you need to get quite close.


The software is on GitHub. The README provides installation instructions. Please also watch the video linked to the README to get a sense of the mechanics.

This software is ready to install.


For each language family, the documentation of available constructs is in the repository’s README. Note that the video shows you how to work with the variants of the strings language.

Here is the list of language families for the whole semester. Check the calendar for out- and in-dates: The arithmetic language is a little tricky. Don’t get demoralized if you can’t figure it out entirely. Remember the grading policy!
  • arithmetic

  • conditionals

  • fun-calls

  • fields

  • mut-vars

  • mut-structs

  • eval-order

In the video, we end up with a small set of programs that can tell all the variants apart. This set is called a classifier. For each of the language families above, your task is to produce a classifier that distinguishes the variants, akin to that in the video.

Observe that a given program in the classifier set may distinguish only one variant from all the others (just as in the video). When there are more than two variants, it is not always easy to produce a single program that can tell them all apart (i.e., produces a different answer in each variant); nor is it necessary. Sometimes a classifier is much clearer to a reader (like a grader!) if each program distinguishes just one variant; then it has a clear purpose.

Therefore, don’t try to get overly clever. You can if you want for your personal satisfaction, but that doesn’t mean you have to turn that in to us! Or if you do, try to also include simpler, less ambitious programs as well. A classifier does not have to be the minimal set necessary to tell apart the variants.