webyrd / minikanren-with-symbolic-constraints

Version of miniKanren I normally use. Has ==, =/=, symbolo, numbero, generalized absento constraints.

miniKanren-with-symbolic-constraints

The version of miniKanren I normally use. Includes ==, =/=, symbolo, numbero, generalized absento constraints.

Good for writing Quine-generating interpreters, etc. :)

Also includes eigen, which represents universally quanitifed variables. Beware: this implementation does not support use of eigen with constraints other than ==.

Also includes multi-query variable version of run. For example, (run (q r s) (== (cons r q) s)).

Running

Chez and Vicare

(load "mk.scm")

Racket

(require "mk.rkt")

Guile

(load "mk-guile.scm")

Chicken

(load "mk-chicken.scm")

Running Tests

After loading miniKanren as above,

(load "test-all.scm")

regardless of scheme implementation.

minikanren-with-symbolic-constraints's Issues

Using miniKanren-with-symbolic-constraints and Petite Chez Scheme

Hello!

Thank you for these files.

Issue:
I am encountering (a) a slight infelicity and (b) an unexpected error and while using these files and Petite Chez Scheme (PCS).

Description of errors with light background:
I am presently reading 'The Reasoned Schemer'. My intention is to use these files and an interactive shell of PCS to type the examples from the book and experiment. I have saved the mini-Kanren-with-symbolic-constraints files in a directory. In a terminal window, I navigate to this directory and open an instance of PCS; then -- in accordance with the markdown file -- I (load "mk.scm") and (load "test-all.scm"). All the tests pass.

(a) the slight infelicity:
When I type "succeed" (i.e. > succeed), the output is #.
When I type "fail" (i.e. > fail), the output is #.
When a user inputs a goal, the output is #.
So, since the output is the same for both "succeed" and "fail" when a user inputs a goal, the user cannot decide using the output whether the goal succeeds or fails.

In order to test whether a goal succeeds or fails, I have been inserting the goal into a cond expression (e.g. (cond [ 'a] [else 'b])). then checking whether the output is a or b.

Is this the standard way of checking whether a goal succeeds or fails?

(b) the unexpected error:
In the book, the first question I encountered that uses conde is question 44. My translation (from the book's typesetting to Scheme) of the expression in that question is:
(conde
(fail succeed)
(else fail))
When I enter this into a PCS shell, I get an error message "Exception: misplaced aux keyword (else c) near line 210, char 23 of mk.scm"

When I enter:
(conde
(fail succeed)
(fail))
I get #, as expected.

Is this second rendering (without else') the appropriate translation? i.e. should I ignoreelse''s in the book when translating from the book to the shell?

Thanks again!

Best wishes,
Andrew

P.S. I am a noob. Sorry if these are questions I should be resolving on my own.