🚀 go-pugleaf

Newsgroups: rec.games.programmer
Subject: Re: Game library...
Summary: 
Expires: 
References: <9207220449.AA68324@hal.gnu.ai.mit.edu> <1992Jul22.131418.19638@wam.umd.edu>
Sender: 
Followup-To: 
Distribution: 
Organization: /etc/organization
Keywords: 

In article <1992Jul22.131418.19638@wam.umd.edu> dmb@wam.umd.edu (David M. Baggett) writes:
>
>You're trying to incite me aren't you?  :)  We've discussed this into
>the ground elsewhere, so I won't go into it at great length again.

I'd be the first to point out that game programmers are above the norm of
professionals, as well.  Not trying to incite anything.

>
>write a C++ compiler you'd find similarly difficult challenges there,
>not to mention similar opportunities for clever optimization.  GCC/G++
>was 100,000 lines of code, last I checked.  At least one game library I
>know of has a similar amount of code.  Part of your task as a
>programmer on such a system is just to keep everything in order so you
>know what the heck is where.  The larger a system gets, the more
>difficult this becomes.  So it's a different set of challenges.
>

Very true.

>>The quality of even the artists and musicians who work with the
>>programmers far exceed the norm.
>
>Part of being a professional is *endurance*.  While the art for something
>like Dungeon Master or Ultima Underworld may look simple enough, realize
>that it consists of literally thousands of shapes.  It's not just a matter
>of "can you draw this at all," but rather, "can you draw this quickly, and
>do the same quality of work 1000 more times?"  If you have a PC, load
>up Commander Keen IV and count how many different shapes you see.  It's
>a staggering job to do art for even a simple game like that.  Game artists
>don't get enough credit.

Quite true.  But don't think there isn't endurance in demo production!
Quite a bit of the time, these demos are produced out of day-night
sessions in order to meet competition deadlines.  If you want to talk
about endurance, you really should check out the Ray of Hope II demo
by Majik 12 on the Amiga [nic.funet.fi /pub/amiga/demos/majik12/roh2]
There is MUCH more artwork in this demo than in the average game.
About 4 fonts including one very large one [80x64, I think] in 16 color
greyscale where each letter is a different picture of it's own.  This was
all done by a demo artist with an Amiga 500 and a Russian B/W TV set.

>
>Just like it takes endurance and perseverance not to quit in the middle
>of your 100,000 line C++ compiler, so does it take endurance and
>perseverance to actually put together and fine-tune a couple hundred or
>thousand game shapes.

I'm sure it does.  And it certainly helps to get paid doing it.  There's
certainly much worse jobs than doing game artwork.
>
>What I see missing from most demos (on the Atari at least) is polish.
>They look like the programers got their awesome hack working and then
>lost interest without bothering to really make the thing look
>professional.  This is not really a criticism so much as an observation;
>demo programmers really *aren't* interested in making something shine
>as much as showing off their latest and greatest hack, right?  I mean,
>once you kill yourself to get 262,000 colors on the screen at once
>without flicker, are you really going to bother making sure your text
>font is nice?

You've not seen many of the later Atari demos.  The latest is the Phaleon
Giga-Demo which boots up with an amazing NeXT login-sequence, contains
a Dungeon Master and Shadow of the Beast 'game', plus about 50 works
from other groups.  There is much more attention to small details in
demos than I see in most games.  

Amiga demos generally have much more polish.  For instance, the Odyssey
demo [French group, Alkatraz] is a 45-minute long movie.  Hardwired
is by two guys who have had their work featured on MTV.  Amiga demos
have gotten to the point where there's less a priority on breaking
some record for polygon faces but on the actual design and character.
In that respect, demos have become a true multimedia artform.  A good
example is Anarchy's In The Kitchen [wuarchive.wustl.edu: 
/systems/amiga/incoming/demos/Anarchy_ITK].  

>
>Again, demo programmers aren't doing the same thing better; they're doing
>a different thing masterfully.
>

It depends on how you look at it.  Really, the only difference between
most game programmers and most demo programmers is that the game programmers
get paid to write routines which work on many machines while demo programmers
win contests by getting the most out of one.  Fortunately, there are much
less differences among Amigas than PCs so they can get away with more.
Quite a few conventional coding tricks have been found by demo coders, too.
How about doing bresenham in 3 lines of 680x0 code?  No branching and no
MUL/DIV?  Just a short table and the addx opcode.  I learned this one
from New Mode/Delta Force.  

>>Don't get me wrong.  I don't have anything against game libraries as
>>long as the programmer is supplied with source and the right to 
>>make improvements on them [and an incentive to make them available
>>as updates].
>
>So you *would* be opposed to such a library without source available?
>I'm not sure how using a library without source code would make any
>difference to you, since you write everything yourself now anyway.  If
>you didn't like some feature of the library you could (in theory) write
>your own routines and use only those parts of the library that you
>found acceptable.  How could it be *worse* to have a library?

I said I don't have anything against the library.  I just wouldn't have a 
need for it.  The reason why I would want source available would not be for
myself, but for the purpose that the person using it might be able to
learn how it works, possibly improve on it, and send the improved version
back to the manufacturer so the library will improve.  Hence, it would
have educational value.  To be honest, as of lately, most of the high quality
games have been programmed in Europe and only distributed by US companies.
Has there EVER been a game like GODS, Turrican 2, WolfChild that didn't
require a '386 to run?  Lucasfilm still puts out great stuff but where
is the technical brilliance of the Atari 8-bit games Ballblazer,
Fractalus, etc..  What happened to the days of John Harris?  Where is he
now, anyways??

>
>>But, the argument for game libraries is that people want to do just the
>>fun stuff [*-i.e. design-*] and none of the hard stuff [*-programming-*]
>>where-as, here are demos where the programmer[s] did all the hard
>>programming and very little game design.
>
>You're assuming that programming a game is harder than designing one.
>I think you've got it backwards.  Sure, you can crank out yet another
>Chop Socky move-to-the-right-and-kill-things game, and in that case the
>art and programming might take a lot longer than the design.  But a
>truly good game design can only be cultivated and refined with a lot of
>work (which might be why we hardly see any good designs in arcades
>these days -- designing is not cost-effective).  I think most
>profesional game programmers would say that the coding is the easy
>part.  Perhaps some of them would like to comment?  (We might ask the
>Atari and Williams folks in rec.games.video as well).

Maybe these arcade game manufacturers should hire people based on talent
and vision instead of just based on a degree in computer science.  I
noticed in jobs.offered a post by Atari Games with 'Knowledge of C
required' and 'Knowledge of 680x0 assembler helpful.'  Aim high, eh?
>
>>Do you think people would go to all this work if hardware programming
>>wasn't fun?  esp. for no pay?
>
>Nope; obviously some people find it fun.  No argument there.  I'm not
>going to try to tell you you're not having fun!
>
>It probably had Atari *pissed*, if you want to know the truth.  "Yet
>another applcation that won't run on the <STe, Mega, TT, Falcon>"  But
>for a NEOchrome slideshow, who cares?  Like I said, a game is a
>different matter.  I think they did us a disservice by using that
>technique in commerical software.

This was before the STe, Mega STe, TT, and Falcon and from the post on
Compuserve, they seemed impressed and bent on finding out how it was done.
It certainly is a disservice to at least not make it known which machines
may have incompatibilities.  It is also surprising that the games that
use this technique don't provide a 'calibration' progam so that when 
you buy the game, you can adjust the timing for your system.  Some software
does this already.
>
>>How does it contradict my opening statement?  I'm confused.
>
>Simply in that your opening statement was that game libraries are
>ridiculous, and that you couldn't imagine anyone using one.  Maybe
>I was reading more into that comment than was intended.
>

Probably were.  It can be hazardous taking anything I say too seriously.  
d:^)


>Dave Baggett
>dmb@wam.umd.edu

STratoHAK/TPPi
psteffn@gnu.ai.mit.edu