From: dododge@wam.umd.edu (David O. Dodge)
Subject: Re: VR: Low-End 3D -- Big in U.K., Why Not the U.S.?
Date: Mon, 14 Oct 1991 06:03:47 GMT
Organization: Workstations at Maryland, University of Maryland, College Park



In article <1991Oct14.034241.11786@milton.u.washington.edu>, trpw@cix.compulink.co.uk (Tim Watson) writes:

> I am a little confused about the "Low End 3D" postings on sci.virtual-worlds.
> 
> I'll explain (in a round about way):
> 
>         A few years back I was involved in writing some 3D software for the
>         PC and on a good day my program could splat about 1000 shaded 
>         polygons per second on the screen of a 25Mhz 386. This program went
>         on to form the basis of Argonaut Software's game Starglider 2.

I must say that the speed of Starglider 2 on my Amiga still amazes me.

          [stuff deleted]

>         The thing about this situation is that almost all the companies and
>         individuals are British. In fact I have NEVER seen a 3D polygon game
>         from the US that is anything like as fast or smooth as any one of the
>         British products. I don't know why this is, other than that we don't

As far as personal computer based programs I definitely agree.

>         have any money in the UK to build accelarator boards or the like, so
>         we have to write fast low level code to run on our cheap PCs.
> 
> So given that there is a growing community of programmers in the UK who can
> write low-level 3D graphics routines for the PC, I find it a little suprising
> that there isn't a similar community in the US. You might imagine that I was
> shocked to see someone suggest using C to write their 3D routines.

I'm not sure if that last sentence is a reference to my posting a while back
about my 3D-glasses code in C or not. In any case the reason I would choose to
use C is for portability concerns. If I were writing for a specific machine and
speed was a necessity I would probably write machine code instead -- or as
close as I could get to it.

In the case of some workstation-class computers writing in assembly is not
always an option. My experience comes mainly from Apollo and Sun machines,
in which case the manufacturer provides C/Pascal/FORTRAN libraries for
doing low-level 2D graphic work. It's generally not a good idea to go much
lower than these libraries, because the hardware on the various node models
can be radically different and the manufacturer isn't always willing to give
up detailed information about how their libraries/hardware work internally.

In the case of PC's I agree that assembly is the probably the best way to go.
In fact the some of the first things I did on a PC was write some 2D graphics
routines in 8088 assembly...

[from a list of other UK 3D writers...]

> David Braben    This guy writes extremely fast 3D programs on obscure
>                 computers such as the Acorn Archimedes.

I've got "VIRUS" by Braben(?) on my Amiga and although it does bog down every
now and then it's still very impressive.

                                            -Dave Dodge/dododge@wam.umd.edu