Question

I’m having an internal dispute about the right way to do something. Maybe you can help 🙂

Let’s say I have the functions

foo(x, y) 
{  
  if (x satisfies condition c1)
    if (y satisfies condition c2)
      bar(x)
  else
    bar(x)

  ...
}

bar(x)
{
  ...
}

An alternative to this would be

foo(x, y) 
{ 
  doBarFlag = false
  if (x satisfies condition c1)
    if (y satisfies condition c2)
      doBarFlag = true
  else
    doBarFlag = true

  if (doBarFlag)
  {
    ...code from bar() goes in here
  }

  ...
}

And another alternative (slight variation of above)

foo(x, y) 
{ 
  doBarFlag = true
  if (x satisfies condition c1)
    if (y DOES NOT satisfy condition c2)
      doBarFlag = false

  if (doBarFlag)
  {
    ...code from bar() goes in here
  }

  ...
}

Assume that bar() is less than ten lines. Which style would you prefer? Why? I’m leaning toward the first example, but I’d like to know what others think.

Answer 1

How about

foo( x, y ) {
  if ( ! c1(x) || c2(y) )
    bar(x)
}

There’s a name for this logic transformation, which I forget…