RE: RE: Space Wars

I do bounding box checks first (this is a must). That's all I had it doing
at first, but there's no way to determine which "part" of the ship got hit,
so I have to check the segments.

Shot to ship checks are just fine. One line (with endpoints at the old shot
and new shot positions) checking against the segments of the ship.

The problem is when both ships get within each others bounding boxes. Then I
have to check each segment of the first ship against each segment of the
second ship. I originally had 9 and 11 segments for a total of 99 checks and
each check requires several multiplies, and since the 6502 doesn't have a
multiply, it takes several hundred lines of instruction to do each check.
(I've since cut the bounding segments down to 6 and 8, it's not as accurate,
but it's a little faster, but still not fast enough)

I had also hoped to eventually write a 4 player space wars, but using my
current collision detection, it would probably bring to game to a halt if
all 4 ships were colliding at the same time.

I've posted this before, but if anybody has a suggestion on how to spead up
accurate collision detection, please let me know.
I'm implementing the following formula for each segment check (re-written in
assembly, of course, for a total of about 300 lines of code. Remeber the
6502 only has add, sub,and shift)

x1,y1 - segment initial position
x2,y2 - segment new position
x3,y3 - last point from ship used to check collision
x4,y4 - new point from ship used to check collision
a1=y2-y1
b1=x1-x2
c1=x2*y1-x1*y2
r1=a1*x3+b1*y3+c1
r2=a1*x4+b1*y4+c1
if r1 & r2 !=0 & have same sign, they don't intersect
a2=y4-y3
b2=x3-x4
c2=x4*y3-x3*y4
r1=a2*x1+b2*y1+c2
r2=a2*x2+b2*y2+c2
if r1 & r2 have same sign, they don't intersect

This works find for shot collision detection (I'm using the old shot
position and the new shot position for the line endpoints), but it's way too
slow for ship to ship.

-jeff

-----Original Message-----
From: Paul Kahler [mailto:phkahler@Oakland.edu]
Sent: Thursday, February 03, 2000 7:36 AM
To: vectorlist@synthcom.com
Subject: Re: VECTOR: RE: Space Wars

> >Ya know, I am wondering why a co-processor is needed for Space War on
Space
> >Duel hardware, since SD manages to do the colision detection, AND keep
all
> >the other objects moving.
>
> Not to answer on Jeff's behalf or anything, but he's doing real
> point-on-line collision checking which is a fair amount of math. Space
Duel

I too am wondering why a co-processor is needed. To say "he's doing real"
hard stuff doesn't justify adding a co-processor. Just don't do such an
elaborate collision detection. Here's a suggestion:

1) Do a bounding box check first - there should not be more than 1 or *2*
bullets within the box at any time. I assume he's doing this already.
2) For each bullet within the box, rotate/translate the bullet into
ship coordinates and check against a table representing the outline
of the ship. Ya, the rotate will take 4 multiplies, but that's gotta
be faster than anything dealing with each line :-)

<ramble>
BTW, as for the scenix. I was a little let down that they cloned the PIC
'54 architecture. I wrote a chess program for the PIC16C63 that just needs
a little more RAM and stack than the scenix. It takes 30-60 seconds to make
a move and can look up to 6 moves ahead. I've been meaning to wire up a
PIC 16C923 with built in LCD controller to run it :-) No doubt this is the
smallest chess engine in the world (written in C yet).
</ramble>

--
 ___   __   _   _  _
|   \ /  \ | | | || |       phkahler@oakland.edu     Engineer/Programmer
|  _/| || || |_| || |__     " What makes someone care so much?
|_|  |_||_| \___/ |____)      for things another man can just ignore. "
-S.H.
---------------------------------------------------------------------------
** To UNSUBSCRIBE from vectorlist, send a message with "UNSUBSCRIBE" in the
** message body to vectorlist-request@synthcom.com. Please direct other
** questions, comments, or problems to neil@synthcom.com.
---------------------------------------------------------------------------
** To UNSUBSCRIBE from vectorlist, send a message with "UNSUBSCRIBE" in the
** message body to vectorlist-request@synthcom.com. Please direct other
** questions, comments, or problems to neil@synthcom.com.