Re: 100 Random Single Variable Linear Equations

Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:

Hi Mike,

Hey Janis =)
 

is my guess correct that you want to create linear equation samples
to be printed (on paper) and solved? Or is it meant as a programming
course sample? - My suggestions depend on being one or the other...

Just a quick project, nothing serious. Really just thinking aloud
and hoping some of you might offer up your thoughts. Really great
suggestions, I think I'll add some of your ideas to it in the next
few days.
 

I've learned linear equations to contain an additive term, as in
  a x + b = c    (i.e.  a*x + b == c  written as Awk expression),
so I'd have expected the operator to be '+' or '-' (not '*' or '/').
(Otherwise, with  a*x * b , you could just calculate a*b first and
a/b respectively, in case of a division a*x / b, before then doing
the single final lhs/rhs operation. The "both sides" procedures
you describe in your introductory comment would be unnecessarily
complicate if you really meant * and / .)

Ahh but I do like the 'spice' of '*' & '/' (see latest iteration below).

But I do wonder about: 5x vs. 5 * x or even (5 * x)... I've read so many
opinions on this matter. If there an offical standard? I dont know.

One older book I have (from 1917!) has 1-2 paragraphs saying 5x without
an intervening * is very bad form & yet, everybody seems to use it, at
least here the USA.

I wonder about the many temporary variables and technical comments;
most don't contribute to legibility or clearness and are unnecessary.
There could be used better naming for the remaining fewer variables.
It could gain from more structuring, like using a 'random' function
for integers to make the random expressions simpler.
Control structure could be simplified, made clearer; do { } while .
Re-iterating over the stored equations is unnecessary, you can just
print them.

I know, more clean ups are needed. But the embedded documentation
ought to be included IMO, though, it is terribly messy...

(I've added code reflecting these suggestions at the end of my post
in case you'd like to pick an idea or two. I've also changed a few
more details, just in case you wonder about any differences to the
original code.)

Yes, certainly, let me study & consider your code & see if I can weave
it into the project. Sounds interesting.
 

 
    function rnd (n)  # n -> 1..n
    {
        return  int(rand() * n) + 1
    }
 
    BEGIN {
        srand()
 
        while (++serial_number <= 100) {
            do {
                opc = rand() < 0.5 ? "+" : "-"  # choose random operator
                equ = sprintf("%d x %c %d = %d", rnd(20), opc, rnd(20),
rnd(50))
            } while (equ in equations_store)    # avoid duplicates
 
            equations_store [equ]  # memorize generated equation
 
            printf("%3d.\t%s\n", serial_number, equ)
        }
    }
 

Yeah I like your thinking, nice & clear. Solid stuff. I'll put some of
this to work during Christmas.

Here's my latest (before I saw your reply). Checkout 'z', sort of like
a gear in a machine yeah? 'n' too. But my imagination is running wild...

BEGIN {
    srand() # seed random number generator

    # keep generating until we have exactly 100 unique equations
    while (u < 100) {
        a = int(rand() * 20) + 1 # random value 1 to 20
        b = int(rand() * 20) + 1 # random value 1 to 20
        c = int(rand() * 50) + 1 # random value 1 to 50
        z = substr("*-/+", (++q % 4) + 1, 1)      # cycle operators
        e = sprintf("%dx %s %d = %d", a, z, b, c) # formatted equation

        # store equation in array if it doesn't already exist
        if (!(e in equ)) {
            equ[e] = 1 # mark element as reserved
            u++        # increment u for each unique equation
        }
    }

    # print equations
    for (j in equ) printf("%03d. %s\n\n\n\n\n\n\n", ++n, j)
}

# eof

--
:wq
Mike Sanders