OpenEuphoria

Re: brain gymnastics

• Posted by "Carl R. White" <cyrek at BIGFOOT.COM> Jun 05, 2000
• 544 views

On Sun, 4 Jun 2000 15:10:33 +0200, Ralf Nieuwenhuijsen <nieuwen at XS4ALL.NL>
wrote:

>> >    integer i
>> >    i =3D 1
>> >    i =3D -i
>> >    i /=3D i+1             [note a]
>> >    +i =3D +i              [note b]
>> >    i *=3D i-1             [note c]
>> >    -i =3D -i
>> >    ? i
>
>integer i
>i =3D 1
>i =3D -1
>i =3D i / ( i + 1 + ( i =3D +i ) )
>i =3D i * ( ( i -1 ) - (i =3D -i) )
>? i
>
>Does precize the same.

Run the two programs as they stand, Ralf. Yours returns '2'. The first
returns '0'... 

I made the same mistake right after I'd figured out the linebreak trick...

I've noticed those who have tried to solve Tor's problem have all been
doing the same thing:

There's a tendency to see (things like):
    a *=3D a + b + c =3D d
      -- Uh-oh.

...as:
    a =3D a * (a + b + (c =3D d))
      -- Multiply 'a' by (a + b) if c !=3D d, and (a + b + 1) if c =3D d.

...when we should be reading it as:
    a =3D a * ((a + b + c) =3D d)
      -- Make 'a' zero if (a + b + c) =3D d, leave it alone otherwise.

Yet in situations like this:
    if a + b + c =3D d then...

...we know exactly what the expression means!

I reckon that this is some kind of "codical" illusion, and also a darn good
argument for using parentheses to make expressions more readable. 

Carl

--
Insert clich=E9 about inserting witty comment here, here.