[boost] Re: Re: Re: Suggestion regarding next() and prior()

From: Fernando Cacciola (fernando_cacciola_at_[hidden])
Date: 2003-12-02 14:53:01

"AlisdairM" <alisdair.meredith_at_[hidden]> escribió en el mensaje
news:Xns9445C4EBA2ED6alisdairmeredithukre_at_80.91.224.252...
> Rob Stewart <stewart_at_[hidden]> wrote in news:200312021902.hB2J2Ri22303
> @entwistle.systems.susq.com:
>
>
> > I've seen you mention this elsewhere, but I don't get it.  Adding
> > the "d" makes the past tense of the verb, so when you call the
> > function, it will have already advanced the iterator, leaving me
> > to wonder what the call actually does.  IOW, how does temporal
> > mechanics fit with advancing by some number of steps? ;-) (My
> > point is that adding the "d" doesn't connote advancing by some
> > number of steps.)
>
> advance is an imperativ, a command, and indicates it will change the state
> of the passed iterator.   advanceD describes the state of what is
returned,
> a different iterator to the one passed.
>
Exactly.
The name is choosen with a focus on the _expression_ that results.

> That said, I prefer the idea of overloading next and prior. <g>
>
> AlisdairM
>
I understand that reusing the names is important.
The only issue I have with the overlading approach is that in my experience
with advanced(), the number of steps is often given as a variable with the
sign somewhat hidden, that is, you don't _lexically_ know if you're moving
forward or backward. In such a context, a call to 'next' with negative
values, or 'prior' with positive values, might be misleading.

Fernando Cacciola
SciSoft