Re: [AD] Example update

[Chris <chris.kcat@xxxxxxxxxx>]

On Sat, 16 Oct 2004 21:34:23 +0200, Elias Pschernig <elias@xxxxxxxxxx> wrote:
> On Sat, 2004-10-16 at 19:53 +0100, Grzegorz Adam Hankiewicz wrote:
> > Attached proposed example update.
> 
> > + *    Fixed
> > + *    point operations are quicker than floating point operations
> > + *    on lower than Pentium hardware.
> 
> Maybe we should formulate this somewhat more vague, like "Fixed point
> operations may be quicker than floating point operations on older
> hardware." I'm no spezialist, but I think floating still at least isn't
> necessarily faster then fixed point.. so it's not always a mistake to
> use fixed on > Pentium.

Yes.. and I also think it depends highly on other code that may be
running, so throwing a blanket statement that floats are faster than
fixed and P2's or above is mis-leading. Mixing float and fixed could
also be beneficial as you could leave the FPU to do the floating point
work while the CPU does fixed/integer work.

I'd just say something like "floating point operations are generally
faster on modern hardware, although fixed point still has benefits of
its own; such as constant decimal accuracy; and could be taken
advantage of if mixed with other floating point operations.

Bob or someone who knows about CPUs would have to clarrify the part
about mixing floating point and integer math, though. I'm not positive
if it's true or not.

> And, maybe this is a good time to replace the "Hit F6, stupid!" by
> "Waiting for F6 key.".

I'd just leave it like it is. I think it's part of Allegro's charm
that it uses "lax" language over the mechanical "Press [button]
now"-type messages.

> Hm, and actually, ESC should always work.. maybe
> just allow ESC along with F6 - I always found the F6 annoying when
> testing exkeyboard.

Still useful to make sure it's working.. but I wouldn't mind ESC being
a abort key if a message is printed out saying so.

- Kitty Cat