http://qs321.pair.com?node_id=14007


in reply to RE: Flipper
in thread Flipper

Interesting idea about throwing away the result of ++. I disagree...

But, why would we care?

I settled on this code because I believe it specifically addresses the large-scale operation we want, which is increment $flip each time, but reset it to zero when it gets to 4. As opposed to: set $flip according to the following formula...

Each time through, we want to increase the value of $flip by 1. The idiomatic way to say that is ++$flip.

What, fundamentally, is the difference between

$flip = ++$flip % 4
and
$flip = ($flip + 1) % 4
(except being three characters longer) when the entire point is to increase $flip by 1?

I guess my question is: why do you think I am throwing away a result which you only accomplish differently? "++" vs. "+ 1" -- they both add one to the value of $flip. See below for the readability benefits of using ++ to add one to a variable.

I like the shorter (by two characters, but hey...) version

($flip += 1) %= 4
but by your argument, you shouldn't like it, either, since it increments $flip, just to "throw the result of the += away." I don't understand your objection.
It is also much closer to being needlessly obfuscated, with all the obvious negatives... (though my coworkers might be shocked to hear that I don't prefer obfuscation :-)

I'm sure you considered, as I did:

for (1..12){ print "$Cheers[++$flip % 4]\n"; }
and rejected it because we want to emphasize the purpose of $flip, which is to cycle among the elements of the array. We assign to $flip to keep its value within the legal bounds. By this we make it obvious (at a glance) to the poor programmer who comes behind us that $flip will only hold values between (in this example) 0 and 3, because those are the only values we want.

I would argue that the preincrement adds a huge "obviousness" value for the aforementioned fellow programmer. ++$flip triggers an immediate (subconscious?) understanding of our intent. Similarly, <nobr>$flip = ++$flip % 4</nobr> is, I believe, immediately obvious in a way that <nobr>($flip += 1) %= 4</nobr> can never be. <nobr>($flip + 1)</nobr> is plain in its intent (make $flip increase by one each time), but not as plain as ++$flip.

Russ

Replies are listed 'Best First'.
RE: RE(2): Flipper (Maintainability vs. personal style?)
by turnstep (Parson) on May 21, 2000 at 20:41 UTC

    > I guess my question is: why do you think I am throwing
    > away a result which you only accomplish differently?
    > "++" vs. "+ 1" -- they both add one to the value
    > of $flip.

    Um, no, they do not. Saying
    $flip + 1;
    does not change in any way the scalar referred to by $flip, while "++$flip" does. Simple as that.

    ===
    What, fundamentally, is the difference between 
      $flip = ++$flip % 4
            and 
      $flip = ($flip + 1) % 4
    ===
    
    The first changes the value of $flip twice, the second, only once. In the first one, the computer is saying:
    1. Increase the value of $flip by one
    2. Modulo 4 it
    3. Assign the result into $flip
    versus:
    1. Compute the value of $flip + 1
    2. Modulo for that (temporary) value
    3. Assign the result into $flip
      turnstep said: Saying
      $flip + 1;
      does not change in any way the scalar referred to by $flip, while "++$flip" does.

      Yes, turnstep, of course the preincrement modifies $flip.

      The question at issue is, "Why do we care about modifying a variable twice?" Simple as that.

      Go back and read the post and respond to the actual issue, please.

      Russ