Topic: That webcrawler game, obscene preference confederate maximizer (Read 14157 times)

Soulwynd · « **Reply #150 on:** August 10, 2011, 08:01:52 pm »

Already have. It also backups whenever I run the server.

Annnnd it has learned to open webpages:

Spoiler (click to show/hide)

Yellow = commands sent to the MOO. The .program bit was all that I needed to program to get it to download webpages.


.program #18:crawler_test
{host, file} = args;
net = open_network_connection(host, 80);
set_connection_option(net, "hold-input", 1);
notify(net, "GET " + file + " HTTP/1.0");
for header in (this.headers)
  notify(net, header[1] + " " + header[2]);
endfor
notify(net, "");
a = 1;
content_length = 0;
counting = 0;
while (a)
  try
    r = read(net);
  except ex (ANY)
    boot_player(net);
    return;
  endtry
  player:tell(r);
  if (!content_length && typeof(r) == STR && index(r, "Content-Length"))
    content_length = toint(r[index(r, ":") + 1..$]);
  endif
  if (!r)
     counting = 1;
  endif
  if (counting)
    if (typeof(r) == STR)
      content_length = content_length - length(r);
    endif
    if (content_length <= 0)
      a = 0;
    endif
  endif
endwhile
boot_player(net);
.

Now programming Base Crawler:crawler_test.  Use "." to end.
0 error(s).
Verb programmed.

eval #18:crawler_test("www.w3.org", "/Protocols/rfc2616/rfc2616-sec9.html")

HTTP/1.1 200 OK
Date: Thu, 11 Aug 2011 00:49:29 GMT
Server: Apache/2
Last-Modified: Wed, 01 Sep 2004 13:24:52 GMT
ETag: "40d7-3e3073913b100"
Accept-Ranges: bytes
Content-Length: 16599
Cache-Control: max-age=21600
Expires: Thu, 11 Aug 2011 06:49:29 GMT
Vary: Accept-Encoding
P3P: policyref="http://www.w3.org/2001/05/P3P/p3p.xml"
Connection: close
Content-Type: text/html; charset=iso-8859-1

<!DOCTYPE html 
     PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
     "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns='http://www.w3.org/1999/xhtml'>
<head><title>HTTP/1.1: Method Definitions</title></head>
<body><address>part of <a rev='Section' href='rfc2616.html'>Hypertext Transfer Protocol -- HTTP/1.1[/url]

RFC 2616 Fielding, et al.</address>
<h2><a id='sec9'>9[/url] Method Definitions</h2>
<p>
   The set of common methods for HTTP/1.1 is defined below. Although
   this set can be expanded, additional methods cannot be assumed to
   share the same semantics for separately extended clients and servers.
</p>
<p>

ect....

Next gotta make it kill html tags and understand the ones that it needs to understand. Then make a word cloud for fun and profeet.

Poltifar · « **Reply #151 on:** August 10, 2011, 08:04:37 pm »

Out of curiosity, what's the difference between your minimal core and the LambdaMOO minimal core?

EDIT: Also, wow, the way you combined the MOO code with a simple HTTP GET is fascinating... I didn't think it would be so simple to get webpages from a MOO, just creating a connection as if it were a player then sending a command through the notify verb, huh.

Soulwynd · « **Reply #152 on:** August 10, 2011, 08:09:52 pm »

No idea. But if it has anything to do with lambdacore it assumes you will be using classic mud rooms and players/rooms come with a bunch of verbs and properties you either don't want to use or have to change.

I started mine from minimal.db to avoid that.

Quote from: Poltifar on August 10, 2011, 08:04:37 pm

EDIT: Also, wow, the way you combined the MOO code with a simple HTTP GET is fascinating... I didn't think it would be so simple to get webpages from a MOO, just creating a connection as if it were a player then sending a command through the notify verb, huh.

You can use the same method to send and receive emails from inside the MOO. Or connect to an IRC network. Or have chats going between two or more MOOs. Or connect to the MSN network. Etc.

You can also assign the connection to a player-like object and have it interpret incoming lines like regular commands.

The only problem I see is that the dns resolve and the connection itself happen in the main loop of the MOO, so if it takes 1 second for it to resolve & connect, everyone lags for one second. Threading it inside of the moo doesn't really solve that either. So you do have to be careful with it. If you spam open_network_connection() commands you can lag everyone. But once it's connected, it's in the queue like everything else, pretty much lag-free.

Ps.: Does the Content-Length html tag includes the headers? Because that website ends with around 400bytes still left to read.

DrPoo · « **Reply #153 on:** August 11, 2011, 12:20:51 am »

Holy shit you are actually making this!
DrPoo stands at the egde of the sane world, looking down into the fractal masses of porn and spam, waiting to take a dive and jump into the directionless world of random and horrors, taking the shape of a banana phone.

MaximumZero · « **Reply #154 on:** August 11, 2011, 12:39:36 am »

Ooookay. As I have done for other games, I will offer you something. My bandwidth. As an American with an absurd computer and asinine amounts of bandwidth, if you need a host, I will be that host. So long as I'm not unknowingly trawling kiddy porn or sex sales, we should be okay. Internet and computer are always on, so that won't be an issue (except in the case of power outage.)

Soulwynd · « **Reply #155 on:** August 11, 2011, 02:12:58 am »

I might take on that offer if this gets somewhere, thanks.

(So knowingly trawling kiddy porn or sex sales would be okay? xP)

counting · « **Reply #156 on:** August 11, 2011, 08:15:59 am »

As I walking home from the grocers, I thought about how to phrase the monsters. And I think it's possible to use agent-based idea and genetic algorithm to let the "monsters" evolved by themselves.

I know you've all probably seen TRON, and the "native programs" are probably some kind of GA programming results. So let's actually do it. If there is rule as [(noun_A) (verb_B) (noun_C)], it can be a sequence of genes. And let the players becomes the "natural forces" to "select" which lives and which dies (in GA, it's fitness function to decide). So over time the populations should reflect what's "best" to be like "real monsters" in players eye. And with mating and mutation functions (probably a "eater" function as well to include unseen none, verb, adj, etc) to adding the verities. I think it can be the future steps without using dictionaries to defined and selected a "monster phrase". (This could possibly used on players creation, but I think there are more to be defined first)

And I got a new icon logo.

Poltifar · « **Reply #157 on:** August 11, 2011, 09:34:48 am »

Quote from: counting on August 11, 2011, 08:15:59 am

As I walking home from the grocers, I thought about how to phrase the monsters. And I think it's possible to use agent-based idea and genetic algorithm to let the "monsters" evolved by themselves.

I know you've all probably seen TRON, and the "native programs" are probably some kind of GA programming results. So let's actually do it. If there is rule as [(noun_A) (verb_B) (noun_C)], it can be a sequence of genes. And let the players becomes the "natural forces" to "select" which lives and which dies (in GA, it's fitness function to decide). So over time the populations should reflect what's "best" to be like "real monsters" in players eye. And with mating and mutation functions (probably a "eater" function as well to include unseen none, verb, adj, etc) to adding the verities. I think it can be the future steps without using dictionaries to defined and selected a "monster phrase". (This could possibly used on players creation, but I think there are more to be defined first)

This would be awesome, though a full genetic algorithm with evolution and reproduction and food and an environment (of sorts) might be too much. How about we just use the idea of using the nouns as a gene sequence? If we use the name of creatures as the seed of the random stat generator for a creature, we would have a system that insured that all creatures with the same name had the exact same stats. For a bit more complex system, have each word in the creature name be the seed for part of the stats, such that players learn to recognize what certain words entail a creature is capable of. And this system could eventually be 'upgraded' to the more complex evolving-creatures if it is deemed feasible.

freeformschooler · « **Reply #158 on:** August 11, 2011, 09:40:32 am »

Soulwynd, do you have a url/ip for the server yet? If not, still awesome. Always following.

DrPoo · « **Reply #159 on:** August 11, 2011, 10:15:20 am »

So the word "Zombie" would define that the thing survives getting their limbs hacked off, and that they prey on intelligent lifeforms?
So a Zombie Parrot Whale would be a flying undead deep-sea creature capable of mimicing its audible enviroment?

I like that, if its rogue like 2d MMO, it would be neccesary to simulate creature behavior all the time in player occupied pages, if a MOO, processing wouldnt matter the slightest, we could have a few thousands of pages running with all flavors of trolls and bill cosbies while we just sit on the B12 page building a fortress pretending to be the dwarfs of the internet!

counting · « **Reply #160 on:** August 11, 2011, 11:37:45 am »

Quote from: Poltifar on August 11, 2011, 09:34:48 am

Quote from: counting on August 11, 2011, 08:15:59 am
As I walking home from the grocers, I thought about how to phrase the monsters. And I think it's possible to use agent-based idea and genetic algorithm to let the "monsters" evolved by themselves.

...

This would be awesome, though a full genetic algorithm with evolution and reproduction and food and an environment (of sorts) might be too much. How about we just use the idea of using the nouns as a gene sequence? If we use the name of creatures as the seed of the random stat generator for a creature, we would have a system that insured that all creatures with the same name had the exact same stats. For a bit more complex system, have each word in the creature name be the seed for part of the stats, such that players learn to recognize what certain words entail a creature is capable of. And this system could eventually be 'upgraded' to the more complex evolving-creatures if it is deemed feasible.

GA doesn't require such settings. It's an algorithm with procedures.

Code: [Select]

Initialization gene pools;   //this could be random or preassigned starting sets 
Do
  growing;   //using genes to build agents/creatures and evaluate
  calculate fitness;   //natural or human selection
  sorting and selections;   //pick the winners/survivors;
  mating and mutating;   //crossover genes(many different methods), and introduce new genes
  repopulate genes/species pool;   //using the previous mating to create next generation
Loop;   
//If the fitness function has a goal, it will end loop. Otherwise it will be constant loop

The gene can be a passive gene (like you said) or active gene (affecting the growing process). The human will do the "human selections" to choose which monsters they will "interact" with, and also picking up the winners. (or by killing the losers, depend on how you select). The processes within loop doesn't need to be linear. At any given time an agent can be at either one of the 5 stages. Most of the time stay in "waiting selections", only when constructors are called will "grow", and only when before destructor will called "mate and reproduce", then the gene is stored in gene pool and waiting to grow into new creatures (like an egg).

Soulwynd · « **Reply #161 on:** August 11, 2011, 01:05:26 pm »

Yes, it will work somewhat like GA, except there will be no algorithm behind it.

Something cannot be a "Zombie" something if there's no mention of Zombies in a webpage. So it's not up to a fitness value or human selection to decide what word combination is best. Plus, I've come to decide, to keep things interesting, that words will not gain fixed stats. So the blue adjective in a page can mean something and something entirely different in another page. When killing monsters and doing stuff, you will gain fragments of the words with a small bit of stat from the original word, so it's your job to put fragments together to get a word out of it. Then later, if you mix blue with blue, you will get a new blue word that would be the breeding product of the other two. But then you should be able to mix blue with sky and get a new phrase "blue sky" that has stuff from both words. If you make another "blue sky" you can mix both together.

I think the only words I will bother creating some more static values for will be colors and sizes. Mostly because it would be nice seeing "blue sky" in the game as "blue sky".

DrPoo · « **Reply #162 on:** August 11, 2011, 01:58:35 pm »

Quote from: Soulwynd on August 11, 2011, 01:05:26 pm

Yes, it will work somewhat like GA, except there will be no algorithm behind it.

Something cannot be a "Zombie" something if there's no mention of Zombies in a webpage. So it's not up to a fitness value or human selection to decide what word combination is best. Plus, I've come to decide, to keep things interesting, that words will not gain fixed stats. So the blue adjective in a page can mean something and something entirely different in another page. When killing monsters and doing stuff, you will gain fragments of the words with a small bit of stat from the original word, so it's your job to put fragments together to get a word out of it. Then later, if you mix blue with blue, you will get a new blue word that would be the breeding product of the other two. But then you should be able to mix blue with sky and get a new phrase "blue sky" that has stuff from both words. If you make another "blue sky" you can mix both together.

I think the only words I will bother creating some more static values for will be colors and sizes. Mostly because it would be nice seeing "blue sky" in the game as "blue sky".

I like that! Cant wait till the game is up so i can try it out and do some wizardry.
I thought of a verb: Wield X like a Y where Y is a name of an utility, such as hammer, sword, knife etc. So the game dosent have to determine how you should use something, however using stuff in an absurd way should increase the chance of things going wrong. SO i could say Wield Toaster like a bat, then i type Fire Baseball at Zombie Clown Poo, then i would smash that baseball at the Zombie Clown Poo, and there would be a pretty high chance of the baseball or the toaster breaking into pieces. Also crafting, would not be Craft X using Y, Z, 1, 2 and 3, but more like Hammer X apart/together with Y, or Weld together/cut apart X using Y with Z fuel, Hammer X into Y with Z etc. And it would use the word property inheritance thing for that.

Soulwynd · « **Reply #163 on:** August 11, 2011, 02:19:46 pm »

I don't think there are hammers, swords, or bats in web pages... Except as words.

I want attacks to be thematic. Like hacking, corrupting code, uploading porn, DDoS, trolling, etc.

DrPoo · « **Reply #164 on:** August 11, 2011, 02:29:36 pm »

I just want a game where everything is random stuff from the internet, anything, and you can do anything with it, wether trolling it, eating it, using it as a sword, anything.
No need to be JUST about the internet.

Also crafting could simply be about breaking things and putting them together. And it could simply just be breaking an item into items wich are the words its made off, so breaking a soda can, would give you a soda and a can, wich cant be broken more, since they are single words, then you can break that bill cosby and put together the can and the cosby and make a cosby can wich you can store fluids in, wich you can then label if you want, be it "Bill Cosby Merchandise - made from real cosbies!" or "Fuck shit penis product".

All you would need is a weapon your wielding, and glue/gaffa tape/poo/*insert item you picked up on a page*
We need some "physics" though, as in stuff breaking, and stuff merging, and stuff being throw etc. I can code that when the game is online.

Bay 12 Games Forum

News:

Author Topic: That webcrawler game, obscene preference confederate maximizer (Read 14157 times)

Soulwynd

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Poltifar

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Soulwynd

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

DrPoo

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

MaximumZero

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Soulwynd

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

counting

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Poltifar

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

freeformschooler

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

DrPoo

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

counting

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Soulwynd

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

DrPoo

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

Soulwynd

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus

DrPoo

Re: That webcrawler game, lilyturf seminationalistic unchariot house: aedoeagus