Checkpoint 2

Complete a partially-implemented OCaml starter — the goto field, failure propagation, and DEFINEd functions — matching Movement II's two sessions

Due: 2 April 2027 · Weight: 20%

The brief#

Complete a starter OCaml implementation so it executes labeled statements with a goto field driven by success and failure, and calls user-defined functions declared with DEFINE — the two things Movement II’s sessions built up over weeks 5–6.

You are starting from Checkpoint 1’s working evaluator, not from a blank page: the starter is Checkpoint 1’s solution with this movement’s new material stubbed out as failwith "TODO" — the goto-field dispatch and DEFINE/call-frame handling are the gaps, everything from Movement I (literals, arithmetic, variables, string concatenation) already works and should not need to change. Your job is to fill in exactly the Movement II gaps, using the starter’s existing statement and evaluator representation rather than redesigning the pipeline around them. Download the starter at checkpoint-2-starter.zip; it includes a sample transcript showing the exact input/output behaviour your completed evaluator should match once every TODO is filled in.

As with Checkpoint 1, this is a correctness checkpoint with a narrow brief. A response that reproduces the sample transcript’s behaviour on the given inputs, and behaves sensibly on inputs of the same shape the transcript didn’t happen to cover (a goto field with both S and F present, a function called with too few or too many arguments, a recursive call at least two levels deep), is a complete response. Anything from Movement III onward — pattern matching, ./$ assignment, the pattern primitives — is out of scope here; it belongs to later checkpoints.

What you submit#

Your completed .ml/.mll source files, in the same project layout the starter arrived in, building cleanly with the starter’s own dune setup. Include a short note (a few sentences is enough) naming any place you departed from the starter’s existing structure and why — this is the “quality of execution” half of the criteria below, and it’s also where you declare any assistance you used, per the course’s academic integrity policy.

The marking: block above renders as a weighted criterion table: the bulk of the mark is whether your evaluator actually reproduces the sample transcript’s behaviour; the remainder is whether your completed code reads as a reasonable continuation of the starter rather than a bolt-on.