Summary: In this laboratory, you will consider mechanisms for verifying the preconditions of procedures. You will also consider some issues in the documentation of such procedures.
a. Open up your library and verify that it includes spot-new, spot-col, spot-row, and spot-color. If you are missing any of those procedures, add them.
b. As you may recall from the reading, we will be checking if values are spots (or at least whether they can be interepreted as spots). You may find the following procedure valuable for checking that precondition. Add it to your library.
;;; Procedure:
;;; spot?
;;; Parameters:
;;; value, a Scheme value
;;; Purpose:
;;; Determine if value represents a spot (that is, was created
;;; by spot-new).
;;; Produces:
;;; is-spot?, a Boolean value
;;; Preconditions:
;;; [Standard]
;;; Postconditions:
;;; If value can reasonably be interpreted as a spot, then
;;; is-spot? is #t.
;;; Otherwise, is-spot? is #f.
(define spot?
(lambda (value)
(and (list? value)
(= (length value) 3)
(integer? (spot-col value))
(integer? (spot-row value))
(or (rgb? (spot-color value))
(color-name? (spot-color value))))))
c. Create another tab or window for doing this lab. At the top of your definitions pane, add a command to load your library.
d. In the reading, we worked with the example of finding the leftmost spot in a list of spots. Add the following code to your definitions pane.
;;; Procedure:
;;; spot-leftmost
;;; Parameters:
;;; spot1, a spot
;;; spot2, a spot
;;; Purpose:
;;; Determine the leftmost of spot1 and spot2.
;;; Produces:
;;; leftmost, a spot.
;;; Preconditions:
;;; spot1 and spot2 have the correct form for spots.
;;; Postconditions:
;;; leftmost is equal to either spot1 or spot2.
;;; leftmost is either in the same column as both spot1 and spot2 or
;;; has the same column as one, and is to the left of the other.
(define spot-leftmost
(lambda (spot1 spot2)
(if (< (spot-col spot1) (spot-col spot2))
spot1
spot2)))
;;; Procedure:
;;; spots-leftmost
;;; Parameters:
;;; spots, a list of spots
;;; Purpose:
;;; Find the leftmost spot in spots
;;; Produces:
;;; leftmost, a spot
;;; Preconditions:
;;; spots is nonempty.
;;; spots contains only spots.
;;; Postconditions:
;;; leftmost is an element of spots
;;; For each spot, s, in spots
;;; (spot-col leftmost) <= (spot-col s)
(define spots-leftmost
(lambda (spots)
(if (or (not (list? spots))
(null? spots)
(not (all-spots? spots)))
(error "spots-leftmost: requires a non-empty list of spots, received"
spots)
(if (null? (cdr spots))
(car spots)
(spot-leftmost (car spots) (spots-leftmost (cdr spots)))))))
You may have noted that spots-leftmost requires an
all-spots? procedure.
Here's a definition.
(define all-spots?
(lambda (lst)
(or (null? lst)
(and (spot? (car lst))
(all-spots? (cdr lst))))))
a. What preconditions should all-spots? have?
b. Is it necessary to test those preconditions? Why or why not?
c. Compare your answers to the those in notes at the end of this lab.
d. Document the all-spots? procedure.
a. Verify that spots-leftmost returns an appropriate
error when given incorrect inputs.
b. In the corresponding
reading, there is an extended version of
spots-leftmost that reports different errors.
Replace your current spots-leftmost with the new
version and verify that it does, in fact, report different errors.
Here is a procedure, index-of, that takes a value,
val, and a list, vals, as its arguments and
returns the index of val in vals.
(define index-of
(lambda (val vals)
; If the value appears first in the list
(if (equal? val (car vals))
; Then its index is 0.
0
; Otherwise, we find the index in the cdr. Since we've
; thrown away the car in finding that index, we need to add 1
; to get its index in the overall list.
(+ 1 (index-of val (cdr vals))))))
And here are some examples of index-of in use.
>(index-of color-red (list color-red color-green color-blue color-yellow))0>(index-of color-blue (list color-red color-green color-blue color-yellow))2
a. What preconditions should index-of have?
b. Arrange for index-of to explicitly signal
an error (by invoking the throw procedure)
if vals is not a list.
c. Arrange for index-of to explicitly signal
an error (by invoking the throw procedure) if
val does not occur at all as an element of
vals.
>(index-of color-black (list color-red color-green color-blue color-yellow))Error: The value does not appear in the list.
d. Some programmers return special values to signal an error to the
caller, rather than throw an error. If val
does not occur as an element of vals, why might
it be better to have index-of return a special
value (such as -1 or #f) rather than throwing an error?
Explain your answer.
e. If val does not occur as an element
of vals, why might be better to have
index-of throw an error?
f. Once you have explained your answers do d and e, you may want to check our notes on this problem.
g. Rewrite index-of using a husk-and-kernel
strategy.
When you're done thinking about these questions, add
index-of to your library as it is a very useful
procedure.
Consider the following procedure that increments the red component
of color by 64.
;;; Procedure:
;;; rgb-much-redder
;;; Parameters:
;;; color, an RGB color
;;; Purpose:
;;; To produce a color that is much redder than color.
;;; Produces:
;;; newcolor, a color
;;; Preconditions:
;;; FORTHCOMING
;;; Postconditions:
;;; (rgb-red new-color) = (+ 64 (rgb-red color))
;;; (rgb-green new-color) = (rgb-green color)
;;; (rgb-blue new-color) = (rgb-blue color)
(define rgb.much-redder
(lambda (color)
(rgb-new (+ 64 (rgb-red color)) (rgb-green color) (rgb-blue color))))
a. What preconditions must be met in order for
rgb-much-redder to meet its postconditions?
b. Should we test those preconditions? Why or why not?
In a number of exercises, we were required to blend two colors. For example, we blended colors in a variety of ways to make interesting images, and we made a color more grey by averaging it with grey. In blending two colors, we are, in essence, creating an average of the two colors, but an average in which each color contributes a different fraction.
For this problem, we might write a procedure,
(rgb-weighted-average ,
that makes a new color, each of whose components is computed by
multiplying the corresponding component of fraction
color1 color2)color1
by fraction and adding that to the result of
multiplying the corresponding component of color2
by (1-fraction). For example, we might compute
the red component with
(+ (* fraction (rgb-red color1)) (* (- 1 fraction) (rgb-red color2)))
a. What preconditions should rgb-weighted-average
have? (Think about restrictions on fraction,
color1, and color2.)
b. How might you formally specify the postconditions for
rgb-weighted-average?
c. Here is a simple definition of
rgb-weighted-average.
(define rgb-weighted-average
(lambda (fraction color1 color2)
(let ((frac2 (- fraction 1)))
(rgb-new (+ (* fraction (rgb-red color1)) (* frac2 (rgb-red color2)))
(+ (* fraction (rgb-green color1)) (* frac2 (rgb-green color2)))
(+ (* fraction (rgb-blue color1)) (* frac2 (rgb-blue color2)))))))
Rewrite rgb-weighted-average to use a husk-and-kernel
strategy to test for preconditions.
Consider a procedure, (, that builds a
new list by substituting list-substitute
lst old
new)new for
old whenever old appears
in lst.
>(list-substitute (list "black" "red" "green" "blue" "black") "black" "white")("white" "red" "green" "blue" "white")>(list-substitute (list "black" "red" "green" "blue" "black") "yellow" "white")("black" "red" "green" "blue" "black")>(list-substitute null "yellow" "white")()
a. Document this procedure, making sure to carefully consider the preconditions.
b. Implement this procedure, making sure to check the preconditions.
Consider a procedure, (, that, given a list of spots and two
colors, makes a new copy of spots-substitute
spots old
new)spots by using the
color new whenever old
appeared in the original spots.
a. What preconditions does this procedure have?
b. Implement this procedure, using the husk-and-kernel structure to
ensure that old and new
are rgb colors and that spots is a list of
spots, before starting the recursion.
index-of, Revisited
Rewrite index-of using tail recursion and a husk
and kernel.
Here are some possible solutions.
a. The all-spots? procedure needs a list as a parameter.
b. It depends on how we will use all-spots?. If we are sure
that it will only be called correctly (e.g., after we've already tested
that the parameter is a list or in a context in which we can prove that
the parameter is list), then it need not check its preconditions. Otherwise,
it should check its preconditions.
d. If index-of explicitly checks its precondition
using member?, we end up duplicating work.
That is, we scan the list once to see if the value is there, and
once to see its index. Even if index-of
does not explicitly check its precondition, the caller may be
called upon to do so, which still duplicates the work. By having
index-of return a special value, we permit the
client to have index-of do both.
e. In some cases, programs should stop when there is no index for a specified value. For example, a program that tries to look up a grade for a student should not continue if the student does not appear in the list. There are also some instances in which careless programmers do not check the return value, which can lead to unpredictable behavior.