On Sat, Aug 31, 2002 at 02:03:47AM +0900, R. Lahaye wrote:
> Are you saying I should do something like:
>
> frnt::OriginType GUI_conversion(int const GUI_choiceItem)
> {
> switch (GUI_choiceItem) {
> case 11: return(frnt::RightBaseline);
> case 10: return(frnt::RightBottom);
> [...]
> case 2: return(frnt::LeftTop);
> case 1:
> default: return(frnt::DefaultOrigin);
> }
> }
What's wrong with that? [Apart from the superflous () in return and the
strange name?] The conversion has a purpose. Name that purpose and call the
function like that.
> And then later in the script I call ONCE the first conversion (apply part):
> igp.origin = GUI_conversion(fl_get_choice(extra_->choice_origin));
Even if functions are called only once they might help in splitting a long
sequence into logical chunks. Of course there is no sense in doing that for
one-liners, but 15 lines look ok to me.
> However, I can't see any advantage of having the trouble of making a
> function, but calling it only once!
You do one thing (conversion) in one line in the "main" function. And if
the function has a suitable name there is not even a need to look up that
function when reading the code.
> Are you still in favour of using
> such help functions in this context?
Yes. A bit at least. But as I am off for ten days now you got to fight with
other peoples now anyway ;-)
Andre'
--
Those who desire to give up Freedom in order to gain Security,
will not have, nor do they deserve, either one. (T. Jefferson)
