Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Topics - Numsgil

Pages: 1 2 [3] 4 5 ... 22
31
Announcements / LaTeX
« on: April 27, 2009, 08:16:38 PM »
[img class=\"tex\" src=\"http://www.forkosh.dreamhost.com/mathtex.cgi?\LaTeX\" alt=\"LaTeX: \LaTeX\" /] is supported now.

[img class=\"tex\" src=\"http://www.forkosh.dreamhost.com/mathtex.cgi?\int_a^b \! f(x) \, dx\" alt=\"LaTeX: \int_a^b \! f(x) \, dx\" /]

Code: [Select]
   [tex]\int_a^b \! f(x) \, dx[/tex]
[img class=\"tex\" src=\"http://www.forkosh.dreamhost.com/mathtex.cgi?\int_a^b \! f(x) \, dx\" alt=\"LaTeX: \int_a^b \! f(x) \, dx\" /]

Code: [Select]
     [math]\int_a^b \! f(x) \, dx[/math]

32
Darwinbots3 / Bots as sailboats
« on: March 12, 2009, 01:38:37 PM »
Just a fun little idea I had.  Wouldn't be the default, but might be fun to play with:

With bots as capsule shapes, you can think of them as boats with a bow and an aft, etc.  Which got me thinking about how exactly a bot is able to mover around.  I've always imagined it's with little retrorockets, or maybe cillia.  But maybe they use sails.  Maybe you can set a wind in the sim to blow in a certain direction, and the bots have to learn to maneuver with their rudder and sail to get around.  They could change the angle of their sail to best catch the wind and use their rudder to navigate and turn.

From a biological standpoints it's pretty silly of course.  But I bet the bots would be the first AI used to steer a wind powered vessel

33
Darwinbots3 / Darwinbots3 Installer
« on: December 06, 2008, 08:27:40 AM »
Announcing the first DB3 installer!  clicky.

It doesn't have anything in it actually.  It just sets up a place on disk and some registry keys.  This is my first experience with an installer, so I'd appreciate people trying it out to check for any weirdness.  I'm going to start work on an installer for the source code now.  It'll probably work out that the source code installer reads the registry keys set up by this installer to find the proper place to install itself.

34
Darwinbots3 / Visual Studio 2008
« on: December 02, 2008, 07:09:38 PM »
I've upgraded to Visual Studio 2008, just to let everyone know.  Mostly I did it to access the built in unit testing capabilities in 2008 (NUnit was giving me a headache).  Anyone wanting to contribute will need 2008 now instead of 2005.  If you go to the wiki, there are instructions at the top for getting the free version.

I'm not sure how XNA and .NET versions interact, so I may or may not upgrade the projects to use .NET 3 or 3.5.

35
Announcements / Leagues
« on: December 02, 2008, 05:23:11 PM »
I've modified the league forums so anyone can post a topic in them now.  Instead of having just one person in charge of running the leagues, now anyone can do it.  Just start a new thread in the appropriate league forum everytime you run the league.  Remember to attach a zip of the bots and the league table so others can pick up where you left off.  Replies aren't allowed in the league forum, if you want to discuss the leagues please do so in the bot tavern (I'll move any threads that people start without realizing).

For the time being everyone should use either 2.43.1k or 2.43.1L (hopefully it doesn't matter which), since the newer versions don't handle leagues well.

F3 still needs a rules post, so someone write one up and I'll pin it.

36
Darwinbots3 / XSLT anyone?
« on: November 23, 2008, 11:21:00 PM »
Does anyone know xslt?  I want to transform the output text of a command line tool, and it lets me specify an xslt transform file to do just that.

37
Announcements / F2 League run
« on: October 01, 2008, 12:27:26 AM »
Commander Keen has been kind enough to run the F2 league over the last few days with the bots that have been piling up in the starting gate.  Check out the current standings here.

38
Bot Tavern / F3 league
« on: September 26, 2008, 03:21:24 AM »
We might want to either extend the league to have more places (maybe 100), or invent a "newbie" league or something.

39
Off Topic / GenePool 6 released
« on: June 02, 2008, 04:25:25 AM »
I got this email just now:

Quote
Dear Friends of GenePool...

Please check out GenePool version 6 - Beta version. It is now available
for the Mac as well as Windows. You can get it here...
http://www.swimbots.com/betaVersion6.html

--------------------------------------
Mailing List
--------------------------------------
Many of you have corresponded with me over the years regarding Darwin
Pond and the various updates of Gene Pool. Thank you so much for your
suggestions, bug reports, and encouragement. I decided to collect all
your emails into one big pile and send out this notice. If you do not
want to be included in the Gene Pool mailing list, please let me know.

--------------------------------------
What's new in version 6?
--------------------------------------
* Most important - yet invisible - is the fact that the entire codebase
was re-written from scratch in C++, and the bulk of development was
shifted over to the Mac. From there I make periodic ports to my Windows
version - which is easy thanks to an API that I built (which, by the way
should make it easier to make a Linux version some day). Please know
that this fundamental re-write will help me add the various features you
have been requesting over the years. So, while you may not see your
requested feature, it should be easier to add now.

* Genetic engineering - you can open up the genes of a swimbot and tweak
them in real-time (don't expect it to be intuitive. Hey - it could be
worse - you could be working for Craig Venter).

* More genetic diversity - richer swimbot colors, and more phase-shifty
motions.

* Improved user interface - no longer looks like it was drawn with a
crayon.

* "Freeze Time" - move stuff around while the clock is stopped.

--------------------------------------
Known Bugs
--------------------------------------
* Sometimes it crashes on Windows for no apparent reason. Fixing this is
my top priority right now. If you have access to a Mac, try it on that -
I (think) it's more stable.

* The new swimbots are more diverse, but a tad sensitive, so the
population dies out more often than in version 5. I will be carefully
tweaking parameters to try to keep populations surviving longer.

* Great Walls not finished yet. (Important? Leave out? Suggests welcome).

* Files incompatible with version 5. If this is important to you, I can
look into this - otherwise, I'd rather not - it's a headache. Also, pool
files don't appear to be working between Mac and Windows - for some
reason. This will be fixed.

--------------------------------------
Unknown Bugs
--------------------------------------
I'm sure there are bugs and design flaws I haven't noticed yet. Please
let me know if anything weird or annoying happens while running GenePool.

--------------------------------------
New Features Still Pending
--------------------------------------
The MANY features you have requested are stashed away in my various
emails, in lists, and in my head. I'll try to add a few cool features
from the long list over the next year.

--------------------------------------
Open Source?
--------------------------------------
At some point in the future, a few select components of GenePool will be
opened for others to contribute code. This might be accompanied by
making Swimbot DNA human-readable and possibly adding a scripting API.

--------------------------------------
Fixes During Beta
--------------------------------------
Periodic fixes and updates will occur during Beta. Before the year ends,
GenePool 6 will be announced done. I hope we can have some productive
feedback during that time. I will set up a blog to facilitate
discussion. Thanks for your help!

JJ Ventrella
http://www.swimbots.com/betaVersion6.html
http://www.ventrella.com

40
Biology / Crows are pretty damn smart
« on: May 23, 2008, 07:23:01 PM »
I was taking my daily walk from my office to the corner and back, when I saw something that amazed me:

A crow had something large in its mouth (either a bread stick or a piece of bologna or something like that), and flew in to an empty Lot.  It placed the food on the ground, and proceeded to dig a small hole by ripping out a few tufts of grass.  It then picked up a piece of the food, placed it in the whole, and covered it back up with the tufts of grass it had pulled out.  It then took the rest of the food and dug a whole a couple of feet from the first, again by ripping up tufts of grass.  And then it buried the rest of its food, again covering it back up with the tufts it had pulled out.  Finished, it flew off.

I don't know if I'm just biased against non-mammals, or if most of the city birds I'm familiar with are pretty live-in-the-moment instead of thinking ahead, but it downright amazed me.  Did some google searching and found this on crow food hording.

I've seen lots of interesting behaviors from the crows on my walk.  Like one time where it looked like 5 or 6 birds were dive-bombing a lone crow in the field to scare him.  They'd pull up at the last second, and the crow they were dive bombing was merrily ignoring all their efforts.

So I guess the moral of the story is that you should never underestimate the intelligence of "pests".

41
Bugs and fixes / Dramatic program size increase
« on: May 17, 2008, 05:06:34 PM »
Just noticed this, and I'm curious what the issue is.  Somewhere between version 2.43j and 2.43v, the program jumped from 1900KB to 9300KB.  It's stayed at the 9300KB level ever since.  What got added that bloated the size of the exe so much?

42
Suggestions / Swimming physics and drag
« on: May 14, 2008, 04:25:26 PM »
I'm playing around with the equations for swimming physics and drag.  I think this is on the top of your todo list, too, Eric.  Anyway, I found this article helpful.  Check out pages 8 to 10.  It talks about this integral:

Drag Force = Int((n dot v) * n) dS over the surface (S) of the body, using the instantaneous velocity (v) of the point (dS) in question and the normal (n) to the surface.  For ties the normal is constant, the instaneous velocity is easily arrived at by lerping the velocities of either bot, and the length is simply bot2.pos - bo1.pos.  So I think we can build an equation that is physically accurate to allow for swimming motions in the present version.  I don't think it will properly take in to account the energy needed to move the tie to produce the motion in the first place, but it's a start.

I'll try to plug it in to Maple when I get home tonight to arrive at a closed form for a drag force from the tie on either bot.

43
Bugs and fixes / League bugs and misc
« on: March 30, 2008, 04:06:32 PM »
Getting crashes when starting a league when the folder does not exist.  Run time error 5.  It's also crashing when it tries to save (or it did once, the previous bug makes it difficult to test this one).

So I've included my DB install.

Also, totally seperate issues: when I add a species, its default color says green when the actual color is random.  Color should say "random".  Also, it would be nice if every time I start new or change, the current settings are saved to disk, so even if there's a crash I don't have to re-enter everything.

44
Suggestions / Leagues
« on: March 29, 2008, 12:27:52 AM »
Suggestions for leagues:
  •  Costs are sort of weird right now, and encourage bots to be built that are simply gross, but still manage to game the system.  Like Guardian.  I'd like to see just costs from DNA length.  Something really modest like 1 nrg per 100 bps per cycle.  Get rid of costs for conditions and stores.  Stuff like Guardian is really a specialist league (conditionless).
  • Also, I'd include a cost per body point, as a way of balancing small and numerous bots (which each get charged for DNA costs) and large behemoth bots.  1 nrg per 1000 body points, or something like that.  We should try to aim for a good balance between DNA costs (anti small bot) and body costs (anti big bot).
  •  Field size, especially with the larger bots of modern versions, is really tight.  I'd like to see it increased to 4, which is the largest field that won't overflow the xpos and ypos sysvars.
  •  Corpse mode enabled, with no decay.  Instantly killing veggies is fine, but the energy shouldn't just disappear.  Attrition shouldn't be that easy.
  •  Waste threshold dropped to something like 1000 or even 100.
  •  Max veggies to 100 or even 1000.
  •  Max velocity allowable (not sure if it's set to this or not).
  •  Light friction.
  •  Drop the cost of slime way down, to, say, .01 as the default.  Leave shell costs alone.  Ideally a 1000 body bot should shoot a shot that costs the shooter the same or greater amount of nrg as it costs the defender to repair the damage to shell.
  •  Slow the decay of slime to about a tenth of what it is now.  (This and the above will require some trial and error to find good values).
  •  Tie formation cost to 100 nrg.
  •  Shot formation cost to 20 nrg.
  •  Get rid of the current SB league and replace with a conditionless league (NC).  Rule being that bots can't have any commands that modify the bool stack except for their initial start command (obviously store commands, etc. not counting).
  •  Get rid of the current MB league.  I just think the program has evolved too far for it to make sense.  Most "MB"s these days are more mass colonies than segmented worm.  Which is a Good Thing, since it follows actual biology rather closer.  It just doesn't make any sense as a competitive league.  The older bots are all lame with the new physics anyway.
  •  All bots that are entered in any league are automatically entered in the F1 league.  This rule hasn't been followed in the past, though it should be.
  •  Rerun all the leagues by taking all eligible bots in the beastiary, in chronological order, and running them.  Rerun leagues every time there's a new version of the program.
  •  Always use the newest version of the program to run leagues.  This means buddy drops.  If the standings change dramatically, that's a good indication of a bug to look for, so it's a good stress test of new versions anyway.  Also, this way, new bots using new features get in the league quicker.
  •  Max league size to at least 100.  We have about 200 bots listed as either F1 or F2, not counting short bots, of which there are several.  I don't know that we want to put every bot in a league, though.  We could use Animal Minimalis as an entrant requirement: the bot must be able to beat Animal Minimalis in a contest to get placed in the league (Animal Minimalis is fairly stupid so this shouldn't be too hard.  We could use any bot someone likes for this, of course.  I just think it provides a good link between the bot tutorials on the wiki and getting a league placement).  Obviously NC and SS leagues wouldn't need a gate keeper like this, though I do think some sort of "is able to survive" requirement would be good.

I'll try to list any others that come to mind.  The current league settings are mostly legacy, and I don't think they reflect a good proving ground for modern bots.

45
Bugs and fixes / IM needs to be periodically reset? OPEN
« on: December 20, 2007, 07:17:25 PM »
After running for a few hours, my Darwinbots program suddenly won't connect to the internet.  Resetting the program fixes this, but then it'll happen again a few hours later.  The first hint that something isn't right in the log says something like "error disconnecting".

Pages: 1 2 [3] 4 5 ... 22