______/\___________________________ __ ________________ ___ /\______

______/\___________________________ __ ________________ ___ /\_______
\____ \ ________ _ _ ______ \ / \| \ ________ | \/ ______/
/ | \ _) \ \_/ \ | \ / \ \ _) \ | \______ \
/ | \ \ | \ | \/ \ \ /~\ \ / \
\_____ /_______/___| /________/\____\_____/_______/_________/________/
===\_____/============|____/========================================[+tZ^]==

DemoNews Issue #81
January 29, 1995 - February 4, 1995

------------

DemoNews is a weekly publication for the demo scene. It is produced at the
Internet FTP site ftp.eng.ufl.edu (HORNET). This newsletter focuses on
many aspects of demos and demo making. Everyone is welcomed to contribute
articles, rumors, and advertisements.

Information about HORNET and DemoNews can be found under /demos/README

Start.of.DemoNews.081,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

SIZE: 52,846 SUBSCRIBERS: Last week: 1220 This week: 1241 Change: +21

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

SECTIONS ARTICLES
---------------- -----------------------------------
General HORNET has been busy!
New Uploads Files recieved at HORNET
NAID Survival Guide
Editorial General Protection Relationships
Code Ctrl-Alt-Delete, Now Its a Game!
Assembly Part 3 (It ain't no party)
BSP Trees
Back Issues How to Get 'em, Descriptions
Closing Comments Quote for the Week, etc.

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<General>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

_____HORNET has been busy!

Here are just a few things I would like to make the readers of DemoNews
aware of:

-ANSI: The ANSI is back on HORNET! After more flames than I could count
(even with my socks off), I have decided to have an area on HORNET where
the "packs" will be stored. Syntax Error (organizer of iCE) is now
HORNET's main coorespondent in this field, and he will be maintaining the
files in this area. However, there will not be an /incoming/ansi
directory, only a /demos/ansi. Most of the previous problems stemmed from
misuse of the /incoming/ansi directory. Syntax Error will keep up to 10
megs of the most current and wanted packs on HORNET. He will be
personally responsible for keeping this directory updated. Also, any
ANSI-related mail will be forwarded to him. You can reach Syntax Error at
[email protected].

-MC3PLAY: MikMak, Rao, Stony, and Air Richter are currently working on a
Music Contest ]I[ player. This will be used by the judges and contestants
alike for playing MC3 entries. Remember, this contest will be officially
announced at NAID on April 15, 1995.

-DNDP: DemoNews Plus for DOS. Ior, Psibelius, and Zoltar are currently
working on a DOS-based version of DemoNews. More than a plain vanilla-
ascii reader, this is a full-fledged diskmag! It will include graphics,
music, additional articles from the DemoNews staff, and best of all, will
be released on a MONTHLY basis. Stay tuned for more details.

-DN/HTML: Jeff (White Noise) / HORNET deserves a lot of appreciation.
In the past couple of weeks, he has made enormous strides in making the
World Wide Web version of DemoNews come to life. As the entire DemoWorld
project becomes more concrete, you can expect quite a bit of coverage of
it in DemoNews.

-SLACKING: I have been very slow in getting two interviews done. The
first is with Stone/Dust, and the second is with Gore/FC. I finally got
the questions out to Stone, but the Gore interview still has yet to take
shape. My apologies to these two individuals as well as the readers of
DemoNews for the delay.

-NEW FACES: Two new individuals have joined the staff of HORNET as
columnists. They are Tom Verbeure and GraveDigger. I met Tom through
Jason Nunn (dee-cug, another DN columnist and friend). Tom is an expert
in the nuiances of hard core Assembler. GraveDigger is another new
addition. Recently he did an incredible review of the Egg music for
DemoNews, and this week he has done a "Survival Guide for NAID". I am
very pleased to have both of these gentlemen furthering the HORNET mission.

-LISTSERVER: There are two people not mentioned very often in this
newsletter who do a lot of work for HORNET. Pi is the TRUE listserver
coordinator for HORNET. Whenever the listserver crashes or goes buggy,
Pi is the guy I count on to bring it back up. In a slighly different area,
Martin M. Pedersen does one of the janitorial jobs here at HORNET. :)
Each week when DemoNews goes out, I recieve lots of bounced mail
notifications. I forward all of these to Martin and he sends me a list
of who I need to (un)subscribe or help out. This is not a fun job and I
really appreciate his help.

-ASPHYXIA TRAINER: Over the past several months, I have been working on
converting Denthor's "VGA Demo Trainer Guide" source code from Pascal to
C++. In the background of the whole thing, FateGrinder and I shoot around
ideas about how to optimize different assembler routines, etc. If you
look at my code hard enough, you're bound to find some of his influence in
there.

I know I'm forgetting something... :)

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<New Uploads>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
NOTE: All locations start with /demos and then their respective sub-
directory. Please note however that the actual base directories
(like /pub/msdos/demos) may differ from mirror to mirror.

Location Filename.Ext Size Description
---------------- ------------ ---- ----------------------------------------

/demos

/alpha/NEW answer .zip 40 Answer by Warlock of Amnesty
dfuse .zip 10 Digital Fuse intro by Uncle Bob/Zion
einstein.zip 9 BBS Intro for Einstein BBS
friends .zip 875 Straight Line Connection for 3S Party 94
greetro .zip 199 Happy New Year by Sunrise
kiddo .zip 132 A.P.E.X Presents an Intro called Kiddo
mciesc .zip 14 BBStro for MCI Escapes by Subsystem
meet95 .arj 708 Something from Sti/Euphoria (GUS)
naidtro .zip 144 The NAID Party Intro (SB only)
southfix.zip 79 BBS Intro for South of Heaven by Omicron
tornado .zip 23 Shocker presents Tornado Intro (no mus)
xmas_dem.zip 658 Christmas Demo by The Paralytic Minds
/diskmags bm9 .zip 580 Blackmail Issue #9 (stuff from TP94)
yahoo04 .zip 286 Yahoo #4, The Hangover Issue
/nets dginfo18.zip 16 Demogroups Interchange InfoPacked v1.8
/news nad_u01 .zip 1428 New Age Demo Database Upgrade: 1

/code
/demosrc
fcsp2src.zip 31 FC's StarPort BBS ][ Source
incosrc .zip 241 Source to Inconexia demo by Iguana
isad2src.zip 20 ASM source for Immortal Syndicate BBS
kuk2src .lzh 159 Source for Pleasure Access BBS
sea_code.zip 33 ASM Source to Seasick / VLA
sqd1_src.zip 21 ASM Source to Squid1.com (BBS Intro)
/graph/lens lenssrc .zip 40 BAS, ASM example of a Lens
/graph/pallette fadecode.zip 30 C,ASM source for fades and pallette rot
palrot .lzh 2 Pallette Rotate by Draeden / VLA
/graph/plasma c_plasma.zip 55 C source code to color plasma
jclplasm.zip 88 ASM,C For NICE color Plasma
/graph/shadebob jeffbobs.lzh 75 PAS source for Shade Bobs
/graph/tutor tut9new .zip 43 Part 9 of Asphxia VGA demo tutor in C++
/pmode dpmispec.arj 33 Protected Mode API For DOS Extended Apps
protect .lzh 15 ASM Protected Mode programming example
/sound fmed101 .zip 147 OPL3 FM Sound Editor for Programmers
gp15-pas.zip 46 PAS source for GUS Modplayer
gusenv .zip 1 ULTRASND environment checking in ASM
gusp15 .lzh 19 ASM source for GUS Modplayer
pps110 .lzh 84 Protracker Playing Source/Josh Jensen
/utils basm10 .zip 297 Basic to ASM Language Translator
frmi!150.exe 38 Flat Real Mode Interface v1.5

/music

/disks f10_pck2.zip Force Ten Pack #2
f10_pck3.zip Force Ten Pack #3
f10_pck4.zip Force Ten Pack #4
rot .zip Rotation by Neophyte mini musicdisk
traxx36 .zip Latest Traxx album
/programs/frntend mplay12 .zip Musicplay 12.0 frontend music program
/programs/misc readcda .zip Read digital data from CD using CDROM
/programs/players radv1_0a.zip ADLIB tracker package (anyone use FM!?)
/songs/s3m belly .zip Belly's Theme by Falcon (FM ADLIB)
cannabis.zip Cannabis sativa by Transee
dnc2trnc.arj Dance to the Trance by Hector
doom-hth.arj Doom and Stuff by Hector
dreary .arj Dreary as all hell by Hector
epi-opus.zip Opuscule by MusicMan/Epinicion
firesirn.arj Firesiren by Hector
gonnadie.arj We're All Gonna Die by Hector
k-udream.zip Unfaithful Dream by Boomer the Bass Pig
messiah .zip Messiah by Transee
mindrave.zip Mind of a Raver by Avatar
mystwat2.zip Myst Water by Hector/DMK
mystwatr.zip Part 2 of the above
nois .zip Nois by Kevin, Fast drums & strange sou
pl_dt .zip Song by Plastique/Dep
pl_heart.zip "
pl_know .zip "
pl_move .zip "
pl_rave .zip "
pl_thund.zip "
pl_uknow.zip "
pl_who .zip "
pl_zany .zip "
renais .zip Renaissance by Null
rh_cyu .zip Song by Red Horizon/Dep
rh_insom.zip "
rh_nw .zip "
rh_ready.zip "
rh_real .zip "
rh_samur.zip "
rh_sea .zip "
rh_tod .zip "
rh_vp2 .zip "
sky_jung.zip Jungle Baby Love by D.J. Skyjump
trn-imh .zip In my Hough by Transee
trn-medi.zip Mental Diary by Transee
wait4u .zip Waiting for You by Falcoln (Adlib)
/songs/xm mftp .arj Music for the People by (see below)
moby-go .arj Moby-Go by Pieter Van Den Veen
/text life .zip Some text about a new group

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<NAID>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

_____Survival Guide by GraveDigger

+ . /\
. + . . / \ . +
_____ 3\ __3____ \__/_______ Official Survival Guide
\__ \3 \ 3\ \ + __ \ \ for the
. / \ \ / _ \ / \ \____ \ . North American International Demo
/ _ \/ /_ \/ \/ 3 \ Competition
. / \____/ \__/ __/__ \ April 15 - 16, 1995
\ / . \____/ \__/ + \______/ww
:::\__/::::::::::::::::::::::::::::::::

Here is a practical(?!) guide to surviving NAID, the upcoming North American
International Demoparty, which is to be held in Quebec on April 15th and
16th. There's still a few months left to prepare yourself for this event,
so I decided to provide this guide for anyone who intends to attend. I hope
you find this guide helpful, and remember to pack light. If you plan to
sleep at the school, there will be a lot of people in one big room, unless
you've reserved a room for a group. Of course, feel free to donate me some
space if you're getting a hotel room.

By the way, the above NAID logo was borrowed from the FILE_ID.DIZ of the
naidtro.zip archive. I couldn't draw a logo of my own, and I liked that one
a lot, so there it is.

Personally, I have never attended a demoparty, though I have some ideas on
how to prepare for it, as I'm more familiar with Canada than I am with
Finland or Denmark. I'm hoping this guide can help everyone have the best
possible time at NAID. This being the first party of this kind on this side
of the puddle, I would like to see the date go down in history, and the
party repeated in years to come. We need to show the Euros that we know how
to party, too!

Please keep in mind that this guide is intended merely as a complement to
your regular luggage checklist. For your convenience, though, a small box
has been placed next to each item so you can check it off as you pack it.
Even though you now have this list, don't forget items like toothbrushes,
toothpaste, deodorant or anti-perspirant... Oh, hell with it. I'll probably
be the first to forget these items anyways. (Hope there's a drugstore in the
vicinity?)
_
|_| 1. First, you will need an English-French dictionary. As you know,
Quebec is the French-speaking region of Canada. Comment vas tu? Tu
habites aux Etats-Unis? Parlez-vous francais? Hmm. You might need to
work on your accent, too.
_
|_| 2. Next, you might want to take a life-jacket. Considering that a large
portion of you will be crossing the St. Lawrence River to get to the
school where NAID is being held, you don't want to take your chances
with those bridges.
_
|_| 3. You will want to have a source of caffeine. Whether it be in the form
of money for soda machines or coffee, or your own beverage supply, it
will be required to stay up after hours and party, at which time you
will probably consume more caffeine, and party some more... and hell,
who really wants to sleep anyways?
_
|_| 4. Be sure to pack your teddy bear if you have one. Yeah, just look at
that poor little guy sitting on your bed. How do you think he feels when
you just leave him there?
_
|_| 5. Bring along unfinished programs and songs and such, and maybe you can
get together with someone else who has a pute and some knowledge, and
work on it with them. In fact, here's a C program that I started which I
need help with. If anyone can assist me, please let me know. (I just
recently took up C as a hobby, so please bear with me.)

-----[cut]-----[blah.c]-----

#include <stdio.h> /* Blah v1.0.0 */
/* 1/04/95 */
void main()
{ printf("The program is now running.");
}

-----[cut]-----[blah.c]-----

I'm not sure what this program is going to be used for, but I'm sure once
it is completed it will be really cool. :)
_
|_| 6. If you bring your own computer to NAID, be sure to lock it up, if at
all possible. At other recent demoparties, especially The Party 94,
equipment was stolen or damaged. This only applies if you're not bringing
your kid brother's CoCo, in which case you should leave it out in plain
sight, and hope it gets stolen. Then, when the culprit is caught,
everyone can point and laugh, and this should prove to be a funny
experience, for all except one person. Hey, win some, lose some, right?
_
|_| 7. You might want your own roll of toilet paper. Hey, you never know...
I can imagine that it sucks to get stuck on the can without toilet paper.
Have you ever hobbled down the hall to your toilet paper supply with your
pants down? Uhh.. not that I ever have, I just don't imagine it being
something any of you would want to get stuck doing. Yeah, that's it.
_
|_| 8. If you have a flashlight, bring it with you. This can be used for
finding the bathroom in the middle of the night, or just for reading
porno magazines in your sleeping bag. For the latter option, this will
let everyone around you locate you, and you can then trade porno
magazines with each other. However, please keep in mind that I do not
read porno magazines, nor do I promote pornography. <cough>
_
|_| 9. If you have a beeper or pager, leave it home if possible. It would be
just plain annoying to be watching a demo when all of a sudden, someone's
beeper goes off, half the crowd turns to find the source of the noise,
and misses the best part of the demo. Though it would be funny to hear a
tune with such a sound in it, or even a phone ringing, and watch everyone
looking for it.
_
|_| 10. Keep a calculator with you and keep track of the exchange rate for
your currency. You don't want to get ripped off when buying stuff or
exchanging your money. And in the event you want to sell something, you
can rip them off if you do it right.
_
|_| 11. You might want to bring some homework with you. You'll most likely
be away from home all weekend, and you might have some extra time during
which you can study. Or if you're really lucky, your books will get
stolen and you won't bother going to class anymore.
_
|_| 12. Bring a portable radio. If you live somewhat far from Quebec, you can
occupy yourself by finding cool radio stations that you'll never be able
to hear again after Sunday. You'll also have something to look forward to
when NAID '96 is planned.
_
|_| 13. Two words: breath mints. :)
_
|_| 14. Grab a flat-head screwdriver to take with you. What for? I don't
know. Wait, hmm. Looks like I have run out of ideas for this list.

I hope this guide becomes a part of your trip to NAID. Print it out, have it
translated into french, sign language, and braille. Share this guide with
others planning to attend. Fax it, snail mail it, tape it to your forehead.
Memorize it backwards. Also, If you have any additions to this above list,
please contact me at the email address below. Anyways, hope to see you there!

GraveDigger [uuDW/CoRE]
[email protected]

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<Editorial>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

Much response was generated by last week's humorous editorial entitled
Codethink(School);. This week, I take a U-turn from humorous to dramatic
(hey, you can't expect me to be in the same mood every week).

One small note: I will make frequent use of the pronoun "they" as being
third-person singular. It is incorrect grammatically to have a sentence
such as "Ask someone a question and THEY will give you an answer". The
sentence should be "Ask someone a question and he or she will give you an
answer". However, I will be using "they" as in the above example rather
than filling this editorial with "he/she" everywhere. Enough said, let's
go on...

_____General Protection Relationships

re-la-tion-ship, noun 1. a connection, association, or involvement.

Without exception, every one of us in the demo scene has two things in
common: we interact with computers and we interact with humans. In
essence, we have relationships with both. Under close inspection, these
relationships can actually reveal an amazing amount about who we are as
individuals and how successfully we deal with others in our lives.

Ask a computer what the sum of two plus two is and it will tell you four.
Ask a human the same question and they will probably raise their eyebrows
and wonder which mediation you've been taking. Ask that same person if
they want to go and see a movie. Mabey they're sick, or busy, or tired, or
just don't want to. If you're lucky they might even say "yes". Ask a
computer if it wants to run a program. It might have difficulty, but that
computer will die, trying to make you happy.

If you think about it, a computer is your ultimate friend. It never tells
you to get some sleep or do your homework. Its always there, waiting
patiently for you to make a request. With games, demos, and various GIFs,
the computer can give you hours and hours of entertainment. Even more,
your computer actually changes with you: it remembers what you have done
in the past and keeps track of what's new in your life.

How many times have you had an argument with someone and gone to your room
to watch Second Reality or play a game of DOOM? In that respect, a
computer can help relieve stress. The computer is always in your corner,
willing to help out with any problem you might have at the time.

Over the years, I have had a lot of difficulty in forming lasting relation-
ships with members of the opposite sex. Part of this stems from my
inability to correctly guess what they "mean" all of the time. Say I ask
woman X if she wants to go and eat at McDonalds. The reply is "well...ok".
Let's just take a few possible interpretations of this answer:
1. "I am tired of eating at McDonalds, can't you think of anywhere else?"
2. "I don't really care where we eat."
3. "While I was saying the word 'well', I was actually trying to think
of some place else we could eat. I couldn't, so I guess McDonalds
is 'ok'."
4. "That was just a silly answer. You know I always love to at at
McDonalds!"
5. "The only reason my answer came out like that was because I was
watching the road for you to make a mistake while driving."
6. "What? Are you trying to tell me I'm FAT?"
The list of guesses goes on and on. A computer only HAS two answers,
yes and no (1 and 0). Its nice not having to play mind games once in a
while.

The computer is a shield between you and reality. It can transport your
mind away from the physical world; away from the problems, the headaches,
the work, and the daily routines that we go through every day. The more I
think about it, I realize that our "Escape" CD was very appropriately
named.

As this article draws to a close, I begin to wonder if its any good or if I
should just delete the whole thing and call it a day. I just asked my
computer if it likes the article and it wouldn't tell me. Oddly enough, I
feel like calling my parents right about now.

-Christopher G. Mann / HORNET January 29, 1995

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<Code>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

_____Ctrl-Alt-Delete, Now Its a Game! by Denthor

Telephone conversation :
<Ring> <Ring>
<Click>
Fanus : Hello, Buys residence.
Denthor : Hi there Fanus, whats up? Can I speak to Piet?
Fanus : Ummm ... he's not in ... ummm ... who is this? I don't recognize
you.
Denthor : Don't worry, it must be my new haircut.
Fanus : Oh. This must be Denthor.

Hi all! I am back after a two week absence from the Demuan List. Today I am
going to introduce you to two new games, which should make sitting at the
computer and listening to music a lot more fun.

The first game is one I have described before in my trainers. It requires
two people and a computer. I like to call it Control-Alt-Delete. Here is
how it works. One person gets the delete button, and the other person gets
the Control and Alt buttons. Player one must then hit delete very quickly,
while player two has to try and hit control and alt at the same time. If
the computer reboots, player two wins, otherwise player one stays at his
station. When player one wins, they swap sides and begin again.

This game can provide hours of enjoyment, and really is great fun in
between really boring lectures about stuff you learned years ago. When I
invented the game, I thought I was doing really badly against Pipsy until I
realised that the computer had frozen.

Note : People have caught on how to play at the local university, and sneak
up and play it on my computer while I am telnetting. Be warned. Also, the
game isn't as much of a challenge when you play it by yourself.

The second game I invented while I was on holiday last week (one of the
reasons I didn't write an article) ... we were down the South Coast, stuck
in a hotel room while it poured with rain outside. The Tugela river
overflowed it's banks, the sand bars dissapeared, and the tan I was hoping
to get never appeared. (I have been programming so much I haven't seen the
sun for the past three months, so when I finally get to go away, it rains,
of course. Way to go, Murphy.)

Anyway, with no computers, no phone, South African TV and no nightlife, we
had no choice but to listen to the radio. After a while of mind numbing
boredom, I invented a new game. I haven't named it yet, but here goes. You
listen to a song on the radio. Every time the singer says a word that is
more then one syllable long, you add a point. You don't count words twice,
or all songs would get a high score.

Most of the songs got between three and five, but we lost count when they
played a Counting Crows song. Ace of Base did surprisingly well, wile some
rap artists didn't score anything. Try it, you'll be amazed.

The PC Games Programming Encyclopedia part two will be out on the 10th of
Febuary, I will make sure that Hornet is one of the first sites to get it.
I hope to have up to number 20 of my trainer series in it.

So, I leave you to pick a game and start playing!

Byeeeee...

Denthor - [email protected]

PS. This no-GUS thing is a Creative Labs conspiracy, isn't it?

_____Assembly Part 3 by Jason Nunn

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
\\\\\\\\[ "Implementation Techniques" - Assembly Part III
\\\\\\\[[ By Jason Nunn
\\\\\[[[[
\\[[[[[[[
____________________________________________________________________

In this issue I will be discussing the basic nuts and bolts on how to
implement an assembly program geared towards a demo. This article is
intended for the C or Pascal programmer who hasn't quite got the confidence
to use a full blown assembly compiler. In the first part, you may remember
me telling you of a friend that has reached a turning point in his coding
development, yet he still won't "take the plunge" because he doesn't have
access to all the nice perks of 3GL's, like sine, cosine, and random
functions and larger precision variables than the chip itself. This issue
will hopefully provide that incentive.

But...., before we do that, I would like to first finish off last article's
talk on optimization. I forgot to include the fast string functions. I'm
not going to waffle on too much about this now, as this article is
dedicated to "assembly techniques". All I will do here is show my results,
and give out some tips.

Same things apply on this run - my machine is a 486-33 ISA, operating in
P-mode, and the lower the number, the faster a given instruction is.

STOSB [209729] MOV [EDI],AL [134226]
INC EDI

STOSD [306805] MOV [EDI],EAX [244208]
INC EDI

LODSB [209729] MOV AL,[ESI] [134226]
INC ESI

STOSD [306805] MOV EAX,[ESI] [244208]
INC ESI

MOVSB [228550] MOV AL,[ESI] [142124]
MOV [EDI],AL
INC ESI
INC EDI

MOVSD [384379] MOV EAX,[ESI] [324112]
MOV [EDI],EAX
INC ESI
INC EDI

Of course, REP operations are faster than their equivalent by about half.
In general, it is better to use MOV's and INC's to perform one off
operations. So if you're coding with these instructions, chances are that
you can get a bit more speed out of your code.

Ok, now on with the main talk. Generally, I won't be going into great depth
as there are plenty of tutorials and manuals on the net that explain the
rank basics of assembly. My role here will be to highlight and familiarize
extrordinary things about coding methods of assembly.

If you've never coded in assembly, then it may pay you to write your
equivalent program in a 3GL first and before converting it over. I guess it
depends on the person. I prefer to implement idea's in straight assembler.
I'm comfortable with the language enough to mumble it in my sleep and their
are no barriers or contingencies like there are in 3GL code. You can also
run into serious problems when converting to your target language, but I'm
sure that there are as many negative points about doing this as they are
positive points.

How to crunch huge numbers
--------------------------

One of the first questions a new demo coder may ask is how he/she could
add, multiply or subtract a number that is larger than the precision of the
chip. Well, this really doesn't apply now, as the standard is 32 bits. This
is ample for nearly all calculations, but for those of you that may want to
perform a 64 bit ADD calculation, this is how you do it:

ADD EAX,ECX
ADC EDX,0

In this example, we don't have a 64 bit register, therefore we must make
two data sources, whether they be registers or memory references to act as
one large register. In our case, EDX and EAX act as one. EAX contains the
least significant data and the EDX contains the most significant data of
our 64 bit number. ECX contains the number we are adding to this 64 bit
concatenated register. The basic idea behind this is that we first add ECX
to EAX. If the number in the EAX register "clocks" then the CPU's carry
flag will be set.

The next instruction - ADC (for those of you that don't know) is a funny
sort of ADD instruction that performs two add instructions. It will first
add the source register to the destination register, and then add 1 to the
source register if the carry is set. Hence the name "ADD ON CARRY". In our
example, if the carry flag is set, we will only add in the carry flag as
the source value is zero. Therefore, if the least significant component of
our 64 bit variable (EAX) clocks, the it will carry over to the EDX
component.

Although the above example only adds a 32 number to the 64 bit number. If
you wanted to add a 64 bit number to a 64 number then you would adopt the
following:

ADD EAX,ECX
ADC EDX,0
ADD EDX,EBX

Where EDX:EAX is the destination 64 register, and EBX:ECX is the source
register.

To add larger precision's, we simply chain!. Here we are adding a 32 bit
number that resides in EAX to a 128 bit number which is stored in
EDX,EBX,ECX and ESI.

ADD EDX,EAX
ADC EBX,0
ADC ECX,0
ADC ESI,0

To subtract, the same principle applies, accept we use SUB and SBB
instructions:

(a) (b)
SUB EAX,ECX SUB EAX,ECX
SBB EDX,0 SBB EDX
SUB EDX,EBX

With the 486's math coprocessor, the large multiplication and division is
more viable than our old conventional way of calculating large numbers;
which as you will see and very slow. Pretty soon, I will be exclusively
using coprocessor calculations in my demos, as they are extremely popular
now. Hence rendering the following code (for me) obsolete. However, for
names sake, I'll discuss the old way of doing things...

For multiplying a 64 bit variable to a 32 bit variable you can use this
algorithm:

MOV EAX,ESI
MUL EBX
PUSH EAX EDX
MOV EAX,ESI
MUL ECX
POP ECX EBX
ADD ECX,EAX

As a formula, the code is equivalent to this: ECX:EBX = ECX:EBX*ESI.

Note that you can chain this one also by taking the EDX value from the
second MUL and multiplying it by the next significant register of the
source and adding that answer into the respective register of the
destination.

Dividing is a little bit more complex. How complex?...this complex:

PROC LONG_DIV
OR EBP,EBX
JZ @@jump_0599
PUSH EBP
MOV EBP,ECX
OR EBX,EBX
PUSHF
JNS @@jump_0548
NOT ECX
NOT EBX
ADD ECX,01
ADC EBX,00
@@jump_0548:
OR EDX,EDX
PUSHF
JNS @@jump_0557
NOT EAX
NOT EDX
ADD EAX,01
ADC EDX,00
@@jump_0557:
MOV ESI,ECX
MOV EDI,EBX
XOR ECX,ECX
XOR EBX,EBX
MOV EBP,0021h
@@jump_0562:
RCL ECX,1
RCL EBX,1
SUB ECX,ESI
SBB EBX,EDI
JNB @@jump_0570
ADD ECX,ESI
ADC EBX,EDI
@@jump_0570:
CMC
RCL EAX,1
RCL EDX,1
DEC EBP
JNZ @@jump_0562
POPF
JNS @@jump_058A
NOT ECX
NOT EBX
ADD ECX,01
ADC EBX,00
POPF
JNS @@jump_058D
JMP @@jump_0597
@@jump_058A:
POPF
JNS @@jump_0597
@@jump_058D:
NOT EAX
NOT EDX
ADD EAX,0001
ADC EDX,00
@@jump_0597:
POP EBP
@@jump_0599:
RET
ENDP

This formula divides EDX:EAX by EBX:ECX. Just in case anybody recognizes
this thing, I've reversed it from a certain popular commercial package (not
giving any names) hehe :). I havn't used it since my real mode days
(which, for the record is about 2 years ago when coding TC669), and it's
basically optimized for that. I've made no attempt to optimize it for
P-mode, as I most likely will never use it ever again.

How to implement a Decimal point (or rather - a hexadecimal point :)
--------------------------------------------------------------------

Now that we have discussed the ways in which we can do a whole range of
calculations, your next question is how to implement floating/none discrete
calculations. For that, we must take a register/memory unit and divide it
into two parts. The number and a mantissa. For the sake of efficiency, you
would typically contain this in a single register, namely a 32 bit
register. I usually use this type of construct (represented in binary):

/------------32 bits------------\
NNNNNNNNNNNNNNNN.MMMMMMMMMMMMMMMM

Here you have a 16 bit actual number, with a 16 bit mantissa. As you can
see the actual number is of a higher order than normal. If to wanted to
extract the number from this variable, you can simply perform a SHR 16.
This will arrive you at the "NNN...." component of the number. Here is an
example of 1 and a half:

0000000000000001 1000000000000000b

If we wanted the discrete part of the number (ie the "1" part), then just
perform a SHR 16, which arrives us at: 0000000000000001. As you can see,
there is no real difference between discrete and non-discrete variables. To
the machine, it's all the same thing. The difference is the way you
interpret the product. Calculations are still no different to normal
numbers. If we wanted to add a "half" to this number then it's as simple
that this:

MOV EAX,00000000000000010000000000000000b ; this is a decimal "1"

ADD EAX,00000000000000001000000000000000b ;this is a decimal "0.5"

So, as you can see, it's not very hard. For multiplication, you're going to
have to include a SHRD instruction, as the number will now be in EDX and
the mantissa in EAX, hence the precision is now larger. This will return
the number back to the EAX 32 bit precision that it should be. Here is an
example:

MUL ECX
SHRD EDX,EAX,16

Here, we multiply EAX by ECX, which arrives at EDX:EAX. Then we just step
down this answer to arrive at the result, witch will now be contained in
EAX. With division, it's the opposite:

MOV EDX,0
SHLD EDX,EAX,16
DIV ECX

Here we are dividing EAX by ECX. Note the preparation just before the
divide.

Signed Data
-----------

As of now, we have only discussed unsigned data. Generally speaking, these
calculations are very simular, but there are some major differences.

Contained in a given 32 register, unsigned numbers go from 0 to FFFFFFFFh,
where as 32 signed data range from 80000000h which is the lowest number and
7FFFFFFFh begin the highest number. When using signed data, there are only
a couple of extra things you must know. Signed data has its own
multiplication and division instructions (ie IMUL and IDIV), and its own
set of conditional jump instructions.

JL (jump if less than) and JLE (jump if less than or equal to)
are a signed equivalent to
JB (jump if below) and JBE (jump if below or equal to).

JG (jump if greater) and JGE (jump if greater than or equal to)
are the signed equivalent to
JA (jump if above) and JAE (jump if above or equal to).

To change our unsigned divider from this....
MOV EDX,0
SHLD EDX,EAX,16
DIV ECX

...To a signed divider, simply substitute the MOV EDX,0 with a CDQ. The CDQ
extends a signed number in EAX into EDX. Example given:

CDQ
SHLD EDX,EAX,16
IDIV ECX

Implementing Complex mathematical relationships
-----------------------------------------------

At one time or another, a coder is going to have to use some sort of
complex mathematical function like triangle ratios, logarithmic factors and
random numbers to implement various things. To create a function that maps
a relationship in real time is basically impossible in efficiently terms.
The only way you can do this is to store relationships in the form of
tables. This may not be apparent to users of compilers like turbo C etc but
electronic calculators, compliers, maths coprocessors, spreadsheets all use
this method of mapping these relationships. it a very fast a convenient way
of doing things.

The first common function is the random function. A random signal can be
achieved using the following algorithm. The product of this function is a
random number stored in the EAX register.

;input: NIL; output: EAX
proc random
mov ebx,[random_seed1]
lea ebx,[ebx*4]
mov eax,[ebx+@@rantable]
mov ebx,[random_seed2]
lea ebx,[ebx*4]
add eax,[ebx+@@rantable]
mov [ebx+@@rantable],eax
inc [byte random_seed1]
and [byte random_seed1],01111b
dec [byte random_seed2]
and [byte random_seed2],01111b
ret
random_seed1
dd 2
random_seed2
dd 13
@@rantable:
dd 0fd8fce7ah,02d7ad7b7h,0f48a8f3ab,04a3b8f8bh
dd 0f2dec542h,0a847fab7h,0f4da81aab,04a348f86h
dd 024547edah,03b535a43h,0b35a535ab,0aa333483h
dd 0fd2f4e7ah,0c525a5b7h,016d3b4a4b,0643b4fd3h
endp

If you expand the table to 256 entries then you could eliminate two
instructions, but there again, it's not worth doing. This random function
will give you a very random signal :). There is only one problem with this
algorithm, and that is, the randomness will always follow the same pattern.
If this feature undesirable, then you may like to make an initiation module
that jumbles up the seeds or the numbers a bit. An obvious way of randomly
choosing a seed, would be to store a fixed reference variable in memory.
For example:

proc randomise
mov al,[043253445h]
mov [byte random_seed1],al
mov al,[012345678h]
mov [byte random_seed2],al
ret
endp

Anyway, I'm going to stop here as it's getting very close the deadline
time. One day, I'll learn not to leave things till last minute. In the next
part, I'll be hopefully finishing up this assembly series and moving on to
my talks of sound/tracker programming (the interesting stuff).

I'll be soon releasing a tracker that I have written called FunkTracker.
With this will be the full source code listing. My discussions will be
based around my knowledge and experience when producing current and past
trackers and players, and discussing implementation and hardware issues. I
also plan to discuss reverse engineering using microsoft CodeView, and plan
to obtain hack docs on the AWE32 card. So this will be all coming up!.
until next time.

See ya
:Jason Nunn

_____BSP Trees by Tom Verbeure

Problem situation: sorting polygons is slow and can be incorrect for
certain view-angles. Heavily influenced by Computer Graphics, Principles
and Practice, I have written this small tutorial for BSP trees, which
solves the problem for static objects and for every view angle.

As I already said: Binary Space Partitioning Tree. Unlike many other
abbreviations, this one really explains a lot of the algorithm: it uses a
tree. It partitions space and it partitions in two parts.

First: it is only usefull in static scenes: no 3D morphing or other goodies
are allowed.

Let's go to the 2D case, 3D is exactly the same.
Take a sample scene:

A\ ----- C|
\ B E |
------ |
|
/ .
/ V
D/

The positive side of the polygons is the side with the defining
character... Ignore V for now.

One could sort this thing during rendering, but as there can be no correct
sort criterium and sorting is slow, we don't want that. Besides, we have
memory to spare :-)

Now, we're going to build a tree that is totally viewpoint independent:

Take polygon B as the root. Polygon B divides space in to parts: the
positive and the negative side (Geee!) We have partitoned space in two.

First, scrap B from the 'not-used' polygons-array and classify the
remaining polygons. Group those on the + side, and those on the - side. As
you can see, polygon C is both on the + and the - side. What to do? Create
2 new polygons C+ and C-, erase C. Is there another complainer ? Nope: all
poly's are on either the + or the - side. Now we have this situation:

B
/ \
A,C+,E D,C-

Not really a tree yet, but we've only started...

Now, do the same thing for the groups at the child nodes, without caring
about those in another child-node.

For the + side of B, we have polys A,C+ and E. Take A as next node polygon.
Neither C+ nor E are on it's negative side (we ignore D and C-). For the
other node, take D as next node polygon. Only C- remains there, and it is
on the negative side. That side of the tree is finished. We have the
following situation:

B
/ \
/ \
A D
\ \
C+,E C-

There's one child with more that one poly left. Take C+ as node polygon, E
become it's child, on the positive side. We're finished.
Situation:

B
/ \
/ \
A D
\ \
C+ C-
/
E

Now, what can we do with it? A lot... Suppose the viewpoint is at position
V. In which order do we have to sort the polygons, when using a back to
front rendering algorithm ? Answer: walk the tree, make sure all nodes
(including children) are visited.

Start at the root. Is V on the positive side? Nope, well, we want the polys
far away first, so walk the positive way. Are we on the positive side of A?
Yep, walk the negative way. It is empty! Ah. Well, draw A first. The go the
positive way. Are we positive of C+? Yep. Negative way of C+ is empty. Draw
C+ poly. Go positive way of C+. E has no child, draw it. Go up until a
non-empty branch is found, draw all node polygon not drawn already. We now
arrive at B again. Draw it. Negative is not visited yet, walk it. We're
negative of D. Positive way is empty. Draw D and go negative. C- has no
child. Draw it. All nodes have been visited. The end.

We have drawn the polygons in following order:

A, C+, E, B, D, C- which is a correct order. The BSP tree has to be
constructed only once and for all. From then on, sorting the polygons is
always correct and in linear time. Standard sorting algorithms can be
proved to be of n*log(n) order of time, so we have an increase in speed as
well.

Disadvantages:

- Memory: one has to have the tree in memory. This can be substantial for
lots of polygons.
- Polygon splitting: one ends up with more split polygons. It is almost
always unavoidable to do splitting.
- Polygons are not allowed to move.

A BSP tree is NOT unique: just pick another polygon as a node and one gets
a different one. In this case, one can avoid splitting polygons: start with
a root and build the following, correct, BSP tree:

C
/
B
/ \
A D
/
E

Tadaam! No polygon splitting!

Building a tree with as little splitting as possible is an exponential of
the number of polygons. As Foley and Van Dam says, just try a limited
number of nodepolygons, pick the one with the least splitting and the tree
will be good enough.

Voila. That's it. Not too difficult I think. Notice BSP trees are also
usefull to sort objects, by using planes that divide the space in such a
way that Object A is on the negative and Object B is on the positive side
of the plane. Very useful (only for non-intersecting objects).

This text is written without the Bible (Computer Graphics, P&P) besides me,
but since I read their chapter about BSP trees many times, it contains
almost the same info.

For polygon splitting algorithmes, there is one in Graphics Gems. I don't
know which one, but buy all four books, you won't be disappointed... :-)

Tom Verbeure
Synergy Design

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<Back Issues>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

_____How to Get 'em

After reading this issue of DemoNews, you may be wondering how you can get
previous ones. Well fear not! There are two different ways to do so:

1: FTP to hornet.eng.ufl.edu and go to /pub/msdos/demos/news/OLD_NEWS and
start downloading anything you see.

2: Now you can request back issues of DemoNews via e-mail. Start a letter
to [email protected] (any subject line) and in the body of the
letter include "get demuan-list <index>" where INDEX refers to the
index number of the issue.

For example: get demuan-list 43

This would retrieve DemoNews #76 (part 1 of 2).

For more recent issues that are split into multiple parts, you must send
an individual request for each index number.

_____Descriptions

Issue Index Date Size Description
----- ----- -------- ------ ----------------------------------------------
75 41,42 12/18/94 68009 A DemoNews Reader, The Birth of Commercial
Life, Editorial: Calm Before the Storm,
Interview with Mello-D, US Demo Scene
(Renaissance meeting), Jelly Tots and Pizza
Shops, Review of Wired '94 Graphics.

76 43,44 12/25/94 92589 Interview with EMF, DemoNews Readers Write,
Kimba's Life Story, X-Mas in the Demo Scene,
CORE, Demo & Music Database, Interview with
Purple Motion/Future Crew, Interview with
Krystall/Astek, Common Sense ][ by Perisoft,
Its X-Mas in Africa, Interview with Maxwood
of Majic 12, Assembly Part ][, Common Sense
Response by Stony.

77 45,46 01/01/95 101100 Chart History, Snowman Near-Disaster, Son of
Snowman, The Party 1994, Making Waves, Using
Assembly Part 2.

78 47-49 01/08/95 111185 The Party 1994: Results and Reviews, Report
by Stony and Friends, What happened to PC-
Demo competition. Editorial: TP94 = ASM94
part 2. Egg2: Trancescrambled Review, More
on Fast Tracker 2.03. General Rambling by
Denthor.

79 51 01/15/95 41832 A Day in the Life of Snowman, Ambient Sample
CD 1, Where's the Sound Blaster, TP94
Graphics review.

80 55 01/22/95 27028 DemoNews/HTML, Traffic Jam, CodeThink(School);
The Solo Sample CD

,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,

<<Closing Comments>>

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

The quote this week comes from "Assembly Language for the PC, Third
Edition". p.174

"A program is never done...but it must be stopped somewhere."

This was intended as a moral for programmers, but with a little rewording
the message is applicable to many areas in life.

See you in CyberSpace,

-Christopher G. Mann (Snowman)-
[email protected]

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,End.of.DemoNews.081.