Description
As we have studied in the class, one of the powerful-yet-simple ways of experimenting with the various choices while designing programming languages is to tweak their interpreters.
In this assignment, we experiment with two important features: lexical vs dynamic scoping and eager vs lazy evaluation.
**Plagiarism warning:**
> Note that this assignment is supposed to be done individually.
> Specifically, you should neither show code to each other nor put it up on web.
> We will be using sophisticated plagiarism checkers and any significant detection will lead to heavy penalization.
> You are free to approach the TAs or the instructor for any queries/doubts.
> Discussions on the mailing list are encouraged.
Part A (7 Marks): Dual Scoping
In this part, you need to extend the provided interpreter kernel to have both lexical and dynamic scoping.
Specifically, a parameter may be declared to be dynamically scoped as:
“`scheme
(define (foo
(lambda (a (dynamic b))
( … ))))
“`
Similarly, a parameter may be dynamically referenced as:
“`scheme
(define (foo
(lambda (a (dynamic b))
(+ a (dynamic-ref b)))))
“`
Referencing a variable that wasn’t declared as dynamically scoped, and referencing a dynamically scoped variable without `dynamic-ref`, should both default to lexical scoping.
Part B (7 Marks): Dual Evaluation Orders
In this part, you need to extend the provided interpreter kernel to have both eager and lazy evaluation.
Specifically, a parameter may be declared to be lazily evaluated as:
“`scheme
(define (foo
(lambda (a (lazy b))
( … ))))
“`
Now `b` should be evaluated lazily, upon need, and `a` should be evaluated eagerly.
Part C (1 Mark): Hotch-Potch!
Explain how could you merge parts A and B (in text), along with expected complications, if any.
Submission
– You need to submit a single zip file named `rollnum-asgn3.zip` (where ‘rollnum’ is your roll number in small letters) consisting of the following three files:
1. `rollnum-parta.scm`
2. `rollnum-partb.scm`
3. `rollnum-partc.txt`
– Make sure the Scheme files contain ` lang sicp` at the top, and that they don’t depend on any other modules.
Evaluation
We will be using a testcase-based automated evaluation to check various standard as well as corner cases. You will get marks based on how many testcases does your submission pass.
