[Skip Navigation] [CSUSB] / [CNS] / [CSE] / [R J Botting] /[CS320 Course Materials] [Text Version] lab/12.html Thu May 10 16:56:52 PDT 2012
Labs: [01] [02] [03] [04] [05] [06] [07] [08] [09] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

Contents


    CS320 Lab 12 LISP Laboratory Number 2

      Check out New things on the Course Web Page

      [ http://cse.csusb.edu/dick/cs320/ ]

      Goal

      This lab should teach you how to create and use new arithmetic functions. We will tackle LISP data structures in the next lab.

      Deliverables

      You should record your insights and confusions in a new cs320 lab page and link it to you home/cs320/index page.

      To earn full credit the work must be done before the end of the lab and should contain a list of at least 3 notes. Each note is a short paragraph with one or two sentences and a new LISP function (or link to a file containing the function). The sentences should say what the function does and what you learned by writing it.

      Let me know by calling me across to your workstation when done.

      Hints


      Net
      1. To run LISP open a terminal window. No source to edit..... in this lab!
      2. To leave XLISP send EOT by holding down CTRL and tapping D
      3. Parentheses (...) are the must important things in LISP.
      4. A comment in XLISP starts with a ";" and runs to the end of the line.
      5. Keep your handout on LISP open beside you and look in it for the rules of LISP.
      6. Open many windows and use your hilight-and-paste feature to save time.
      7. Use an editor to write your functions and their test cases then either
      8. Any test editor is good for LISP, but vi can help you track the parentheses.
      9. Put XLISP code in files that end ".lsp".
        1. In 'vi' you can track parentheses if you do the following
           		:set showmatch
          • Copy and paste into a window running XLisp to run them...
          • or
          • Save the version ( :w in vi) and run xlisp on it. (:!xlisp % in vi)

      10. Some browser may panic if you link to a ".lsp" file. If you put source code in a ".txt" file every browser in the world will read it and your user's will love you.

      (End of Net)

      Review

      Start by rereading your last published page of notes!

      Here are some simple LISP expressions/commands. Load the LISP interpreter and input each in turn. Try to predict what each will return as a value before inputting it:

       		()
       		(+ 1 2)
       		(1 + 2)
       		(* (+ 1 2) (+ 3 4))
       		(+ 1 2 3 4)
       		(A B C)
       		'(A B C)
       		A
       		'A
       		(A)
      If you get one wrong... you may need to go back to [ 11.html ] again.

      Creating new functions

      Here is a function with no arguments:
       		( define  (a) 4321)
       		(a)
       		a
       		'a
       		(a 1)
      Copy and paste the above XLISP commands into XLISP in a terminal window.

      Then define and test a new function called answer that returns the value 42.

      Case sensitive guile

      In Guile/Scheme 'DEFINE' can not be used -- use 'define' instead. XLISP allows both and I use it to high-light built in functions/forms.

      Functions with a single parameter

      Here is how we would define a LISP function with a single argument that returns a list:
       		(DEFINE (square x)           (* x x))
      Here is how you test it...
       		(square 3)
       		(square 4)
       		(square 5)

      Here is how you can use it:

       		(square (+ 1 2 3))
       		(+ (square  3) (square 4) )
       		(+ 3 (square 3) )
      Test the above!

      Here is how XLISP can list it:

       		square
      XLISP does not let you edit a function however!

      Do not leave LISP until you complete the next two steps.

      Here is a function for the cube:

       		(DEFINE (cube x)           (* x (square x)))
      Test it.

      Define a function called fourth that returns the fourth power of a number. Use the fact that the fourth power on n is the square of the square of n:

    1. n^4 = (n^2)^2.

      Test it. And save it...

      Functions with two parameters

      Functions of more arguments/parameters are done the same way using the syntax
    2. ( DEFINE (name w x y z ...) expression)

      Here is a definition of a function with two arguments in LISP:

       		(DEFINE (pythagoras x y) (+ (square x) (square y)))
       		(pythagoras 3 4)
      Some common errors
       		(pythagoras 4)
       		(pythagoras 1 3 4)
       		(pythagoras )
      In our XLISP the value of the function name is an expression defining a function:
       		pythagoras

      Here is a function that return the larger of two expressions:

       		(define (max2 a b) (if (> a b) a b))
       		(max2 1 17)
       		(max2 17 1)
       		(max2 (max2 3 5) (max2 4 1))
       		(max2 (square 3) (cube 2))

      Define a function called min2 that returns the smallest of two arguments. Test it and save it....

      A Binary Search

      Here is a function that searches for the positive square root of a positive number. It uses the square function. It has four arguments:
      1. target: the number for which you need a root. Must be >=0.
      2. lo: a number that is less than the root.
      3. hi: a number that is greater than the root.
      4. error: a number indicating the desired accuracy of the approximation.

       (define (binroot target lo hi error )
       	(let (( mid (/ (+ lo hi) 2.0))) ; this saves the time to recalculate mid
       	 (if (<= (- hi lo) error)
       	     mid
       	     (if (< (square mid) target)
       	         (binroot target mid hi  error)
       	         (binroot target lo  mid error)
                    )
       	 )
       	)
       )
      Here is how I tested it:
       		(binroot 50 0 100 0.05)
      Test it further and trace it.

      The above algorithm will find roots of any monotonic increasing function. Modify it to find cube roots and fourth roots of positive numbers.

      Optional experiments if you have time

        Higher powers

        The following function works out the value of x to the power n when n is a whole number greater than or equal to 0.
         		(define (power x n)
         		   (cond
         			((= n 0) 1)
         			((= n 1) x)
         			(T (* x (power x (- n 1))))
         		   )
         		)
        Here are two test cases
         		(power 2 3)
         		(power 3 2)
        Test with the trace function....

        However this is not a very fast way to calculate powers. There is another one based on these facts:
        Net

        1. If n is 0 then power(x,n) = 1.
        2. If n is 1 then power(x,n) = x.
        3. If n is even then n=2*k for some k and so power(x,n)=square(power(x,k)).
        4. If n is odd then n=2*k+1 for some k and so power(x,n)=x*square(power(x,k)).

        (End of Net)

        We have LISP function EVENP and ODDP to detect parity and (/ n 2) works out a k for us by rounding down.

        Can you figure out how to speed up the original power function?

        Classic factorial

        We now define a new function, command and kind of expression that calculates the factorial of a positive integer. The factorial of the number 0 is 1, and for all other positive integers n you work out n times the factorial of n-1. In LISP we say:
         		(define (factorial n)
         		 (if (= n 0) 1 (* n ( factorial ( - n 1 ) ))
         		 )
         		)

        Here you print out the definition of factorial:

         		factorial
        Here you input expressions that apply factorial to simple numbers:
         		(factorial 0)
         		(factorial 3)
         		(factorial 5)

        Trace how it works:

         		(trace factorial)
         		(factorial 0)
         		(factorial 3)
         		(factorial 5)
         		(untrace factorial)

        Below you can use factorial is more complicated ways:

         		(factorial (factorial 3))
         		(setq n 3)
         		(factorial n)
         		(setq f (factorial n))
         		f
         		(setq f (factorial f))
         		f

        The problem with this factorial is that n! quite a small n is too large to be represented as an integer. Worse n! when n<0 is infinitely small! As a result we should define a safe factorial that returns a string "error" when n<0 or n is too big.

        Program and test this one.

        What does this function calculate

        [ fibo.lsp ]

        An example of top-down design in LISP

        [ primes.html ]

      . . . . . . . . . ( end of section Optional experiments if you have time) <<Contents | End>>

      Leave LISP

      To exit lisp, input the EOT character CTRL/D

    . . . . . . . . . ( end of section CS320 Lab 12 LISP Laboratory Number 2) <<Contents | End>>

    Check the Preparation for next class

    [ ../13.html ]

End