Re: A86: Re: Strategy Games

[Prev][Next][Index][Thread]

Re: A86: Re: Strategy Games


>The screen size is too small, there is not enough memory, and the input is
>inferior.
>RTS games need a mouse, trackball, or pen to be played effectively. Also AI
>is difficult to implement on multiple objects and path-finding is insanely
>slow on a 6mhz processor.
>Indeed, its difficult enough to have simple 5 character sprites moving on
>screen at the same time on a simple scroller.
>
>But I'm sure it can be done to some limited degree. I recommend using
>grayscale. It doubles the memory
>requirements but big deal. Just RLE compress the map and images and you can
>stuff quite a bit in strings
>and Ramp page 1. For 4-way RTS algorithms I recommend you use Kirk Meyer's
>algorithm for 4-way grayscale scrolling maps.
>
>Or you can use my source code, which is somewhat older but it works. Feel
>free to take whatever you want.
>
>http://www.acz.org/3programs.zip
>
>- Matt Johnson

Well, my intent was NOT for an RTS, as obviously a 6Mhz Z80 and a 
128x64 LCD display are not good enough, let alone the problem with 
the interface. However, if you look at RoboSport, it achieves an 
almost RTS-like feel with very little requirements. Also, since I 
plan on making it turn-based rather than RTS, I doubt the user will 
have problems waiting a few seconds for the computer to take its 
turn. On the greyscale note, I will wait until I get the current 
engine done, as I said. This is because I am not 100% certain how 
much space I will have (even though the map data is compact, when put 
into game format, it actually becomes 4x its normal size) after all 
the AI/etc. are put into effect. The AI will be rather simplistic, 
operating on 3 phases: determine behavior ratios on enemy actions and 
recent friendly actions (dmg/attack/kill/etc.) and assign each unit a 
purpose; then go through their units making the unit do its purpose 
(or try to move there); and then phase 3 is build any new units it 
believes is needed, based on behavior. An AI player that is currently 
in an 'expanding' mode will build more building units, rather than 
one that is currently in a defensive/offensive position, where 
workers are needed, but assault units are required in larger numbers. 
This method, coupled with turn-based, should allow me to keep the AI 
data down to a minimum (1 Byte per AI unit, 3-10 Bytes for the entire 
faction).

My whole reason for this is to see if I can pull off a game (even if 
it isn't RTS, compact, etc) that isn't really covered territory. If 
space permits, I may allow a campaign mode where you try to conquer 
all of a 'global' setting, and save the game along the way... of 
course those campaign variables will be large though.

-Krev
-- 
----
"Science describes how our universe reacts in a mathematical 
representation of it, but can be described in regular terms. These 
terms seem to infringe on Religion's role of describing the creation, 
and once Science finds a clue to the creation, Religion declares 
Science a heretic. Religion scolds Science, telling it to stick to 
doing what it does best. To this, Science laughs and complies, going 
after the creation once again. Wouldn't it be fair to listen to the 
fact that these humans could have been a product of the Creator just 
as we are, Borak?" - Cyander, Day for the Xenith: Xenith Dawn 

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
References: