John Hatpin <RemoveThisjfhopkin@[EMAIL PROTECTED]
> wrote:
> huey.callison@[EMAIL PROTECTED]
wrote:
> >John Hatpin <RemoveThisjfhopkin@[EMAIL PROTECTED]
> wrote:
> >> OK, I've just run a test that simply tests the whole sim, with an
> >> accurate timer called "counting seconds in my head". Each 255
> >> 'row' of cells takes about one second, so it would take about four
> >> minutes to test the whole sim. That's a bit long for this
> >> application, not to mention for me developing the damned thing,
> >> with repeated re-runs.
> > I think you're doing it wrong. You've got this invisible thing that
> > makes bird noises, and it loads in a place you control. Why not use
> > that thing as a database cursor?
> > (algorithm deleted)
> > This breaks down if the shape is irregular (like an L, or pretty
> > much anything else) since it's only going to find one rectangular
> > space inside that, but once you've stored that in an array, you can
> > walk randomly within that space, but the initial 'finding of the
> > borders' can be done by the bird itself, can't it?
> Well, during the discovery stage there's no reason at all for the boid
> to actually fly - it can sit at home requesting owner****p or anywhere.
Yeah, but... ...why not?
> And, in fact, the calculations won't be done by the flying object -
> there's a controller object with buttons and displays and **** that
> does the UI and acts as shepherd to its flock of birds, telling them
> where to go.
Oh. That's why not.
No self-controlled bots, huh? Bummer.
> But that's by the by, since the edge-walking algorithm can be equally
> used from the comfort of your own X,Y. In fact, that's a lot better
> too, since moving objects around takes time.
I'm reminded of an old Young Ones episode, where the pilot and copilot
of a commercial airliner are talking to each other.
"My, I hope we don't have a crash."
"They say it's safer than crossing the street!"
"...but we have to do that too!"
and then the plane crashes.
Yes, moving objects around takes time. But don't you have to do that too?
> The problem is, you're doing fine at finding coordinates at the edge,
> but we're looking for the largest contiguous rectangle, not the same
> as the edge when the plot itself is not rectangular.
I'm not finding the edge. I'm finding the edge of a rectangle (and not
necessarily the largest contiguous rectangle) inside the bounds of the
edge, depending on the shape you're in, the direction you start in, and
the direction you turn. And if there's a hole in the middle of it,
somebody else's turf completely bounded by your turf, it won't find that
either.
> Crude maybe, but the simplicity's im****tant because there are other
> complications such as not wanting to fly too close to people on the
> land - nothing worse than a blackbird coming out of nowhere and
> shouting right in your ear.
If you don't want to calculate owner****p dynamically, how do you
calculate where people are dynamically?
> Besides, probably 95% of plots are indeed rectangular, and most people
> with irregular plots already accept compromises. For example, my
> Spanish now-ex-cyber-lover and I used to rent for our home an L-shaped
> plot that was formed by two rectangular plots being conjoined. I
> had to buy and fit two separate, networked intruder alert systems to
> cover the land, not because of its size but because of its shape.
"networked intruder alert systems"? You have virtual burglars?
> You shrug and you do it.
The hell you say.
--
Huey
|
