Author Topic: Anyone out there?  (Read 8111 times)

Offline abyaly

  • Bot Destroyer
  • ***
  • Posts: 363
    • View Profile
Anyone out there?
« Reply #15 on: March 15, 2009, 08:04:52 AM »
That's fine if the groupings are configurable. Then there is room to play with it and see what types of groupings lead to interesting results. Consistent groupings matter less than interesting results, and we don't know which will yield those until we try them.
Lancre operated on the feudal system, which was to say, everyone feuded all
the time and handed on the fight to their descendants.
        -- (Terry Pratchett, Carpe Jugulum)

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7742
    • View Profile
Anyone out there?
« Reply #16 on: March 15, 2009, 08:10:05 AM »
Oh, so you mean like user configurable for the groupings?

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
Anyone out there?
« Reply #17 on: March 16, 2009, 01:52:26 AM »
maybe mutations should be split up into two categories: smooth mutation, where advances are made in small, progressive steps, and sudden mutation, where the changes are totally random. Changing add to sub or sgn to cond for example is a sudden mutation, as changing operators/operator types completely alters the function of the code, as opposed to changing 50 .repro to 55 .repro or .aimdx to .aimsx, where the changes are subtle, yet still noticeable.
...well, relatively speaking.  
« Last Edit: March 16, 2009, 01:55:49 AM by bacillus »
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
Anyone out there?
« Reply #18 on: March 16, 2009, 02:09:52 AM »
Is the new shape of bots still two semi-circles and a connecting polygon? Because I had a good idea for a new kind of bot tie: instead of having the bots stick on the surface of other bots, or being connected by 'string', the endpoint circles are used as 'hinges', so the circles' centers are at the same point. A force on either one of the circles would loosen this connection, so chunks could be 'bitten off' other multibots without the prey slipping out due to gaps in the MB.
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7742
    • View Profile
Anyone out there?
« Reply #19 on: March 16, 2009, 10:37:36 AM »
Hmm, sounds like a good idea XD

The pivot point will probably be at the end points instead of the center of the circle.  I'll just half-disable collision code between joined shapes so that the pivot joined parts don't self collide, but the other end points can.

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
Anyone out there?
« Reply #20 on: March 18, 2009, 11:34:26 PM »
Obviously that's not a new idea. What else isn't? I think it'd be good to get a summary of features up.
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline Prsn828

  • Bot Destroyer
  • ***
  • Posts: 139
    • View Profile
Anyone out there?
« Reply #21 on: March 19, 2009, 03:54:21 PM »
Quote from: bacillus
Obviously that's not a new idea. What else isn't? I think it'd be good to get a summary of features up.

Actually, as of today the new shape of the bots is back under scrutiny.
This is because of my realizing that it is, as far as Nums and I are aware, impossible to make a capsule with varying sizes for the ends without either 1) using complex trig to calculate exactly where the polygon connection goes, or 2) allowing the bots to have a concave (indent like) feature.
Both of these could pose a problem for anyone that wants the program to run at a reasonable speed.

Right now there are several alternatives being considered:

A) Polybots, or bots made of polygons.
 Distorted Elliptical bots, like a bent circle.
C) A combination of A and B, or of A and plain, circular bots like we have now.
D) Capsules can be used, because someone was (hopefully this is the case) kind enough to point out a better solution to the problem with this kind of bot.

As far as the current progress is concerned, we are working on the physics engine, but due to this little oversight about bot shape, are unable to continue until this problem is sorted out.  Of course, other things are being done right now, but this is where most of the action is.
So, what will it be? Will you submit to my will, or must I bend reality to suit my needs?
Better answer before I do BOTH!

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
Anyone out there?
« Reply #22 on: March 20, 2009, 05:06:45 PM »
If the circular ends are the same size, then a line can be drawn between the two. Because this is then an axis of symmetry, the line can be offset perpendicular to itself by X units, where X is the radius. This approach is fairly simple (I've written an arc-poly hybrid collision engine and found that this works every time), but without bringing in basic trig, limits it to equal-sized circles.

I just had an idea for tie info sharing. In addition to using the .inx/.outx variables, a bot could enable a 'synapse' cell component that automatically transmits information from a certain input tie to all its other ties. Although this means information is unidirectional and can cause interesting problems in circuit structures, this shouldn't matter too much, as it models real life better than now.
« Last Edit: March 20, 2009, 05:11:58 PM by bacillus »
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline Prsn828

  • Bot Destroyer
  • ***
  • Posts: 139
    • View Profile
Anyone out there?
« Reply #23 on: March 23, 2009, 03:03:13 PM »
Building on your information transfer idea, perhaps certain locations in the bot's memory could be linked.  For instance, a bot might be able to designate a certain place in its memory as read-only, and send that memory value to another bot it is tied to and store it in a value that can only be modified externally by the other bot.  That was a bit confusing for an explanation, but basically it boils down to the fact that information can only be transferred one-way since two bots could try to modify the same value at once otherwise.

As far as bot shape, it has been decided that we will use Trapezoids for the shape, which is basically the same except for the rounded ends.
A later version may allow for the capsules, but for now this is how it is going to be done.
Also, this made me realize that not only can they be trapezoids, but they can also be triangles if one of the bases is of length zero.
So, what will it be? Will you submit to my will, or must I bend reality to suit my needs?
Better answer before I do BOTH!