Path: chuka.playstation.co.uk!news
From: "Scott Ward" <scott.ward6@virgin.net>
Newsgroups: scee.yaroze.programming.2d_graphics
Subject: Negative gradients...
Date: Fri, 24 Sep 1999 15:33:30 +0100
Organization: PlayStation Net Yaroze (SCEE)
Lines: 23
Message-ID: <7sgeqc$6vn1@chuka.playstation.co.uk>
NNTP-Posting-Host: m219-mp1-cvx1a.hud.ntl.com
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2314.1300
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300

I've just set up a collision routine made up of standard bounding lines and
stuff with look-up tables for co_ord's, gradients, etc.
Surprisingly, for me, I got it working to a reasonable degree of accuracy
without too many hitches but one thing really has me puzzled. Whenever my
algorithm tests for collision between any diagonal line (sprite) and a line
with negative gradient one '-1' (scenery) the accuracy is a few pixels out.
More so the further up ('up' on screen) the line the collision happens. I've
been using fixed point maths for the gradient calculations so all the other
lines work out fine accuracy-wise and there's no problem with lines of
gradient '+1'.
As far as I can tell all other lines with a negative gradient work okay, it
only happens when it's '-1' which mathetically shouldn't decrease the
accuracy at all 'cos it's all integer calculations.

This one's really got me stumped, any help would be much appreciated.

Cheers,

Scott

P.S. I've checked all the look-up tables and all the data's right.