arianne-devel — Development discussions

Re: [Arianne-devel] hrmmm....

[Ben B. <ben...@rm...>]

do you guys have any openGL programmers? I'll be taking a course or two
on it this year, so I will be able to help out with that, and there's a
friend of mine whose specialty is CG programming. He'll most likely want
to get in on the game. The game looks like it has a hella lot of
potential, but I still need to play around with it to get it working
with my machine.  From what I can see, the race descriptions look
perhaps a little too complex, and the level of AI is going to fall short
of the desired level.... but with time it should be possible to get
fairly close.

Ben

disculpe, comprendo espanol, pero hablo muy poco :)



Ibis Fernandez wrote:
> 
> If theres anithing I can do as far as arwork and I nimation I would be happy
> to help out with the project. Im not a programer but I have plenty of
> experience with character animation for TV and and the web.
> 
> ----- Original Message -----
> From: "MIGUEL ANGEL BLANCH LARDIN" <x51...@fe...>
> To: <ari...@li...>
> Sent: Thursday, July 26, 2001 5:11 PM
> Subject: Re: [Arianne-devel] hrmmm....
> 
> > Mensaje citado por: Ben Brownlee <ben...@rm...>:
> >
> > > Anyone ever seen this?
> > >
> > >
> > >
> > > bash-2.05$ arianne-client
> > >
> > > arianne-client: error while loading shared libraries: cannot open shared
> > > object file: cannot load shared object file: No such file or directory
> >
> > It looks like SDL.so is not on the path for the libraries.
> > Try to add it ( I think it is something related to LD_LIBRARY_PATH ) or
> simply
> > install SDL on the /usr loaction.
> >
> > > so apparently it is looking for nothing and can't find it... .. . .
> > >
> > > ./configure && make && make install  all went well and the
> > > arianne-server and
> > > arianne-client are both in the path, so I'm not really sure what else
> > > there is
> > > that is needed (and neither the FAQ, INSTALL-LINUX or README cover this)
> >
> > If this fix it, we can it to the FAQ.
> >
> > Regards,
> >
> >                   Miguel Angel Blanch Lardin
> >
> >              -- http://www.arianne.cx -- Arianne --
> > The free open source massively multiplayer online role playing game
> >
> > nuclear cia fbi spy password code encrypt president bomb iran irak korea
> cuba
> > Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna
> >
> >                     -- Echelon must die --
> >
> > _______________________________________________
> > Arianne-devel mailing list
> > Ari...@li...
> > http://lists.sourceforge.net/lists/listinfo/arianne-devel
> 
> _______________________________________________
> Arianne-devel mailing list
> Ari...@li...
> http://lists.sourceforge.net/lists/listinfo/arianne-devel

Re: [Arianne-devel] hrmmm....

[Ibis F. <too...@sw...>]

If theres anithing I can do as far as arwork and I nimation I would be happy
to help out with the project. Im not a programer but I have plenty of
experience with character animation for TV and and the web.




----- Original Message -----
From: "MIGUEL ANGEL BLANCH LARDIN" <x51...@fe...>
To: <ari...@li...>
Sent: Thursday, July 26, 2001 5:11 PM
Subject: Re: [Arianne-devel] hrmmm....


> Mensaje citado por: Ben Brownlee <ben...@rm...>:
>
> > Anyone ever seen this?
> >
> >
> >
> > bash-2.05$ arianne-client
> >
> > arianne-client: error while loading shared libraries: cannot open shared
> > object file: cannot load shared object file: No such file or directory
>
> It looks like SDL.so is not on the path for the libraries.
> Try to add it ( I think it is something related to LD_LIBRARY_PATH ) or
simply
> install SDL on the /usr loaction.
>
> > so apparently it is looking for nothing and can't find it... .. . .
> >
> > ./configure && make && make install  all went well and the
> > arianne-server and
> > arianne-client are both in the path, so I'm not really sure what else
> > there is
> > that is needed (and neither the FAQ, INSTALL-LINUX or README cover this)
>
> If this fix it, we can it to the FAQ.
>
> Regards,
>
>                   Miguel Angel Blanch Lardin
>
>              -- http://www.arianne.cx -- Arianne --
> The free open source massively multiplayer online role playing game
>
> nuclear cia fbi spy password code encrypt president bomb iran irak korea
cuba
> Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna
>
>                     -- Echelon must die --
>
> _______________________________________________
> Arianne-devel mailing list
> Ari...@li...
> http://lists.sourceforge.net/lists/listinfo/arianne-devel

Re: [Arianne-devel] hrmmm....

[MIGUEL A. B. L. <x51...@fe...>]

Mensaje citado por: Ben Brownlee <ben...@rm...>:

> Anyone ever seen this?
> 
> 
> 
> bash-2.05$ arianne-client
> 
> arianne-client: error while loading shared libraries: cannot open shared
> object file: cannot load shared object file: No such file or directory

It looks like SDL.so is not on the path for the libraries.
Try to add it ( I think it is something related to LD_LIBRARY_PATH ) or simply 
install SDL on the /usr loaction.

> so apparently it is looking for nothing and can't find it... .. . .
> 
> ./configure && make && make install  all went well and the
> arianne-server and
> arianne-client are both in the path, so I'm not really sure what else
> there is
> that is needed (and neither the FAQ, INSTALL-LINUX or README cover this)

If this fix it, we can it to the FAQ.

Regards,

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] hrmmm....

[Ben B. <ben...@rm...>]

Anyone ever seen this?



bash-2.05$ arianne-client

arianne-client: error while loading shared libraries: cannot open shared object file: cannot load shared object file: No such file or directory





so apparently it is looking for nothing and can't find it... .. . .

./configure && make && make install  all went well and the arianne-server and
arianne-client are both in the path, so I'm not really sure what else there is
that is needed (and neither the FAQ, INSTALL-LINUX or README cover this)

I'm running Slackware 8.0 on an x86 with kernel version 2.4.5

Thanks to anyone who might know what this is

Ben

Re: [Arianne-devel] intro

[MIGUEL A. B. L. <x51...@fe...>]

> 	I plan on helping out with the development of the game. I know java,
> C/C++ and asm to varying degrees. I'm an honours computer science
> student at Royal Military College in Canada, and have a decent
> background in AI (workin on a neural net driven fingerprint recognition
> system) and have some crypto experience. I'm just starting to take apart
> the source and find out what makes the game tick (so far) and will
> likely just lurk around and play the game and patch small problems as I
> find them. However, if there are a bunch of small tasks that need to be
> done (coding tasks plz) that anyone knows of, just drop me a line (even
> if it is to yell at me an tell me to "find the fux0ring things
> yourself!! ; )

Welcome Ben,

Well, by now there is no such small task on Arianne. You can chase bugs or help 
on pieces of the project that you found interesting.

We are near a release ( Yes, we are :-) so all our work is now to add all the 
code features ( GUI & RP ) and stabilize and clean it.

There are several big task empty:
- AI
- Object System
- RP Rules ( in Python )

Just, try to get confortable here and then choose yourself what do you want to 
do.

Regards,

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

Re: [Arianne-devel] Introductions and questions

[MIGUEL A. B. L. <x51...@fe...>]

Mensaje citado por: Paul Bohme <ar...@sp...>:

> Hey there
>   I like the ideas behind Arianne, and there seems to be some good
> momentum 
> at this point.  I have varying amounts of spare time (the Day Job goes 
> between spans of being boring to eating me alive) but might have a few 
> neurons to throw into the mix.  Have been playing and coding on ROM MUDs
> for 
> some years, so may be useful.  I've been using C/C++ both as a hobby and
> 
> professionally for 10+ years but fear I'm going mushy with all of my
> current 
> work in Java.  ;-)


Hi, welcome to arianne.

>   I grabbed the cvs sources and ran through the general amusement of
> making 
> sure all of the SDL libs were installed and whatnot.  Everything builds,
> installs (more or less properly - my 'config --prefix=/usr' seems to
> have 
> prompted it to put its configs in /usr/etc/?) and the server comes up
> quite 
> happily.  The problem is when any of the clients try to connect it
> croaks 
> with the following output:
> 
> App   :(INFO)-> Initializing Client
> App   :(INFO)-> Connecting to localhost:3124
> App   :(INFO)-> Creating Client connection
> App   :(INFO)-> Loading default Renderer.
> SDL TopDown Renderer:Initializing the TopDown Renderer
> Loading: /usr/share/Arianne/Renderers/TopDown/Main_0_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Land_0_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_0_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_0_1.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_1_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_1_1.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_2_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_2_1.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_3_0.png
> Loading: /usr/share/Arianne/Renderers/TopDown/Human_3_1.png
> App   :(ERROR)-> Serializer.cpp:46 - class Serailizer: trying to read
> beyond 
> end of data
> App   :(ERROR)-> Serializer.cpp:46 - class Serailizer: trying to read
> beyond 
> end of data
> App   :(INFO)-> Shutdowning down Client

Uhm! I have to test code again. Last time I compile it was June... 
 
> Since it sounds like the error is coming from a very common place, it
> doesn't 
> tell me all that much about where to start hunting to look for a
> foul-up.  
> Any hints or pointers?

k, Let's us take it down. 
BTW What is the date of the CVS checkout?
 
> Also, something that generally helps me grok lots of new code reasonably
> 
> rapidly is to run it through Doxygen and see how it looks.  If anyone
> else is 
> interested I can probably put the results up somewhere for all to use.

Well, most of the comments are ready for doxygen.
If you want to complete the work, it is welcome.

BTW if you want a CVS account, talk with Dan following http://www.arianne.cx/cvs 
 guideliness. Also please join Sourceforge.

Regards,

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] intro

[Ben B. <ben...@rm...>]

G'day,

	I plan on helping out with the development of the game. I know java,
C/C++ and asm to varying degrees. I'm an honours computer science
student at Royal Military College in Canada, and have a decent
background in AI (workin on a neural net driven fingerprint recognition
system) and have some crypto experience. I'm just starting to take apart
the source and find out what makes the game tick (so far) and will
likely just lurk around and play the game and patch small problems as I
find them. However, if there are a bunch of small tasks that need to be
done (coding tasks plz) that anyone knows of, just drop me a line (even
if it is to yell at me an tell me to "find the fux0ring things
yourself!! ; )

Thanks,

Ben Brownlee
ben...@rm...
RMC

[Arianne-devel] Introductions and questions

[Paul B. <ar...@sp...>]

Hey there
  I like the ideas behind Arianne, and there seems to be some good momentum 
at this point.  I have varying amounts of spare time (the Day Job goes 
between spans of being boring to eating me alive) but might have a few 
neurons to throw into the mix.  Have been playing and coding on ROM MUDs for 
some years, so may be useful.  I've been using C/C++ both as a hobby and 
professionally for 10+ years but fear I'm going mushy with all of my current 
work in Java.  ;-)
  I grabbed the cvs sources and ran through the general amusement of making 
sure all of the SDL libs were installed and whatnot.  Everything builds, 
installs (more or less properly - my 'config --prefix=/usr' seems to have 
prompted it to put its configs in /usr/etc/?) and the server comes up quite 
happily.  The problem is when any of the clients try to connect it croaks 
with the following output:

App   :(INFO)-> Initializing Client
App   :(INFO)-> Connecting to localhost:3124
App   :(INFO)-> Creating Client connection
App   :(INFO)-> Loading default Renderer.
SDL TopDown Renderer:Initializing the TopDown Renderer
Loading: /usr/share/Arianne/Renderers/TopDown/Main_0_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Land_0_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_0_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_0_1.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_1_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_1_1.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_2_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_2_1.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_3_0.png
Loading: /usr/share/Arianne/Renderers/TopDown/Human_3_1.png
App   :(ERROR)-> Serializer.cpp:46 - class Serailizer: trying to read beyond 
end of data
App   :(ERROR)-> Serializer.cpp:46 - class Serailizer: trying to read beyond 
end of data
App   :(INFO)-> Shutdowning down Client

Since it sounds like the error is coming from a very common place, it doesn't 
tell me all that much about where to start hunting to look for a foul-up.  
Any hints or pointers?

Also, something that generally helps me grok lots of new code reasonably 
rapidly is to run it through Doxygen and see how it looks.  If anyone else is 
interested I can probably put the results up somewhere for all to use.

  Thanks!

  -P

-- 
Paul Bohme - Geek Iconoclast
ar...@sp... - http://www.sponheim.org
Programming today is a race between software engineers striving to build 
bigger and better idiot-proof programs and the universe trying to produce 
bigger and better idiots.  So far, the universe is winning.  - Richard Cook

[Arianne-devel] World development level [was - Re: Group question, Vol 1 #14 - 2 msgs]

[Euan M. <lu...@us...>]

On 30 Jun 2001, at 14:31, ONUR TUGCU wrote:
> ...Starting everything from scratch is fun.
> 
> Maybe we should do the world like that as well. Start
> from scratch. Caves, cave people, then agricultural
> communities, guilds, then cities, rivalries, wars, and
> wow... Anyway it wont take people 4000 years, since
> they know already what they have to do. But then, we
> will need plain, non-adventuring people there who
> plays just for the sake of RP forever. Or is that why
> we need computer generated AIs?
> Maybe even this might be fun. Since there will be so
> many people around.. 
> Just another bizarre idea.

I think it's a good 'un.  As I've said before, I think we ought to 
have continents with stone-age folk, a long, long way away 
from more the more advanced, civilised cultures.  That means 
players can join a 'start-up' continent, or a 'ready-made'.  And 
Arianne would already hacve developed its own  momentum  
by the time any of the continental societies crossed paths.

Plus, I think that there should be some unmapped 'Here be 
dragons' type areas on the map to start of with, that allow us 
to add sections into the world once we've got the hang of how 
it all works.

After all, until the global climate/weather models start up, we 
don't need to know about the mountain ranges, seas, oceans  
and jungles.

Maybe later on we could even backstep the global weathjer & 
climate models to tell us what geography we need to fill the 
unmapped areas with, to end up with the climate that the 
player-inhabited areas have.


Cheers,
     Euan
xl...@us...

'I would live all my life in nonchalance and insouciance,
Were it not for making a living, which is rather a nouciance'
 - Ogden Nash

Re: [Arianne-devel] Re: Group question, Vol 1 #14 - 2 msgs

[<x51...@fe...>]

> So. Are we coding individually, or did someone start
> writing something different about idndividual AI? 
> If so, please direct me to the code. Maybe I could be
> helpful. If not, I will work on whatever I have, which
> seems to be getting better.
> Although I am not so familiar with C, I can follow the
> C instructions. 

As far as I know ( I am the project leader :-) no one is working on AI.
Just because AI is tagged for next release perhaps... but it would be great to 
start work right now. 

Before code we have agreed on Arianne to do designs, so you will need to write a 
small doc explaining what you are going to do and how. If you understand UML you 
can also use it.

AI is going to be VERY related to RP, so you should be in contact with RP code.
 
> Maybe we should do the world like that as well. Start
> from scratch. Caves, cave people, then agricultural
> communities, guilds, then cities, rivalries, wars, and
> wow... Anyway it wont take people 4000 years, since
> they know already what they have to do. But then, we
> will need plain, non-adventuring people there who
> plays just for the sake of RP forever. Or is that why
> we need computer generated AIs?

Yes, We need AI that make the world work.
But we also need that that AI can change goal and start to play the game as a 
NPC.

I will be very happy to have just AI that run roles in ALPHA 3.

> If it was up to me, I would put some monsters in the
> game. For example, one single randomly killing chaotic
> evil vampire AI. Wouldn't that be fun? It would put
> people together good or bad.
> But then, we will need a moderator DM for each unit
> area. Otherwise people would kill eachother just
> because they can. The moderator sets clues to people
> about any kind of massacre so they have a chance to
> find out the person.

My opinion is the one of 0 intervention on the world, it is up to players to fix 
players problems. But there are more opinions on Arianne. 

Anyway I would love that AI don't depend at all on human intervention.


                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] Re: Group question, Vol 1 #14 - 2 msgs

[ONUR T. <hal...@ya...>]

EPITAPH VII

So. Are we coding individually, or did someone start
writing something different about idndividual AI? 
If so, please direct me to the code. Maybe I could be
helpful. If not, I will work on whatever I have, which
seems to be getting better.
Although I am not so familiar with C, I can follow the
C instructions. 

Fuzzy logic was indeed interesting. Now it keeps me
thinking. I am developing a fuzzy logic sentence
analyzer. I hope it works. Starting everything from
scratch is fun.

Maybe we should do the world like that as well. Start
from scratch. Caves, cave people, then agricultural
communities, guilds, then cities, rivalries, wars, and
wow... Anyway it wont take people 4000 years, since
they know already what they have to do. But then, we
will need plain, non-adventuring people there who
plays just for the sake of RP forever. Or is that why
we need computer generated AIs?
Maybe even this might be fun. Since there will be so
many people around.. 
Just another bizarre idea.
If it was up to me, I would put some monsters in the
game. For example, one single randomly killing chaotic
evil vampire AI. Wouldn't that be fun? It would put
people together good or bad.
But then, we will need a moderator DM for each unit
area. Otherwise people would kill eachother just
because they can. The moderator sets clues to people
about any kind of massacre so they have a chance to
find out the person.



__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail
http://personal.mail.yahoo.com/

[Arianne-devel] Fwd: Re: [Arianne-general] mysql++ on Linuxppc

[masahiro m. <el...@aa...>]

sorry i did it again; i posted to general list;
forward the message to the devel list --- sorry for the bother;

>Hi,
>
>I must apologize --- I'd been off line these days:(
>
>And my adventure with mysql wasn't a successful one;
>Something is really messed up.
>I can use mysql (without problem, i guess ), but i have trouble
>with connecting database through C/C++ API calls.
>
>Slow learner as I am , still i'll continue to detect what's going on here.
>Please wait for my good news !
>
>At 11:56 PM +0200 01.6.26, x51...@fe... wrote:
>>  > (I'd show you, Miguel, my mods, maybe tomorrow. I need to sleep
>>>  today....)
>>
>>k, perfect.
>
>P.S. I've been chatting with brian about his new RP codes, and found out that
>my Python on LinuxPPC has something wrong and couldn't test his codes.
>Anyway --- I'll try whatever I can;)
>--
>masahiro minami <el...@aa...>
>
>PowerBook G4/400
>PowerMac G3/233
>MacOS 9.1--MacOS X Public Beta---LinuxPPC
>(Eudora 4.3.2J)
>exception handling : www2.age.ne.jp/~except
>
>_______________________________________________
>Arianne-general mailing list
>Ari...@li...
>http://lists.sourceforge.net/lists/listinfo/arianne-general

-- 
masahiro minami <el...@aa...>

PowerBook G4/400
PowerMac G3/233
MacOS 9.1--MacOS X Public Beta---LinuxPPC
(Eudora 4.3.2J)
exception handling : www2.age.ne.jp/~except

Re: [Arianne-devel] Re: Lack of knowledge, Vol 1 #13 - 1 msg

[<x51...@fe...>]

> EPITAPH-VII
> 
> Thanks to mr K for feedback. 
> 
> to argue about AI at all, I need to make a research on
> fuzzy logic and its possible applications to the
> project. 

Fuzzy logic can be useful to describe human concepts as big, a lot, small, and 
so... so that we can work in the system with these terms... really Fuzzy logic 
is very simple.

Just get the description for the basic operators.

I would also read something about temporal logic, that can be useful too.


> >certain goal is solved(cross the water without
> >getting wet) but how do 
> >you 
> >tell this unexpected goal to the system ?
> >Very interesting to think about imho.
> 
> indeed it is interesting, and I was thinking about it.
> I came to a final decision that this links to a basic
> information. An instinct, if I may so call it, of
> keeping the personal condition or even in general the
> social status, on its best if possible.
> 
> To generalize, one could say, part of the AI is
> continuously checking the important environmental
> variables which have direct, or close, but indirect
> effects on it, and reacting accordingly to this.
> 
> This reaction will depend on the "personality". A set
> of variables and constants which attempt to define how
> the character would act in any kind of situation. 
> The idea of a certain scoring system comes in handy
> here for deciding.
> 
> I think this is as general as it gets. feedbacks?

Yes, perfect.
But AI must build that info from Perception messages just as the rest of the 
players...

> How to link pieces of knowledge? now that is an
> interesting thing to think about. In my opinion,
> again, knowledge pieces should be identical in
> structural definition. 
> Shortly, they should be composed of a usage and a
> condition, which may interact with the AI. 

Well several options appears.
1) First you have to build metaknowledge from Perceptions.
2) Then use that to create knowledge that is useful for the system
We can use:
- Frames
- Expert Systems
- Case Based Reasonners

I don't know very well Frames.
But Expert Systems requiere an expert to build the ruleset, so it will be a bit 
limited.

Case Based Reasoners try to build based on existing examples... so it will take 
out creativity.

CBR is faster than ES. 
We could try a mix of Frames and the other approach.

> Any other thoughts?
> 
> finding the most efficient {means} of reaching a goal
> gains importance here.
> 
> Example: 
> a horseshoe which is being made. It is in very hot
> condition. The goal is already set. The AI needs to
> hold it and change its place. The personality will
> most likely try to avoid long term (in here, quite
> short) contact through limbs with the hot condition
> material, because it is in the harming subset. Seek
> improvement of the condition by searching for
> efficient objects to add into the "using" pool other
> than its limbs. The AI will not be able to make a
> machine to move the material perhaps, but a very
> simple tool in the vincinity could be used to reach
> the goal.
> 
> Wow. Even this simple action looks scary. A tool using
> computer...

Frames can be very useful for this.
But it will requiere a ES or a CBR for later reasoning.

> Another relaxing example:
> Goal: reach from one destination to another.
> AI should be able to check, if there is a tool (a
> horse, a train) in the vincinity which it can use in
> order to reach to the destination more effectively. 

The point is that AI breaks a complex goal in a subset of smaller easier goals 
that can be accomplished by the AI, there is a scheduler called POP that do 
this in a great way.

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] Re: Lack of knowledge, Vol 1 #13 - 1 msg

[ONUR T. <hal...@ya...>]

EPITAPH-VII

Thanks to mr K for feedback. 

to argue about AI at all, I need to make a research on
fuzzy logic and its possible applications to the
project. 

But one thing I would like to add;

>certain goal is solved(cross the water without
>getting wet) but how do 
>you 
>tell this unexpected goal to the system ?
>Very interesting to think about imho.

indeed it is interesting, and I was thinking about it.
I came to a final decision that this links to a basic
information. An instinct, if I may so call it, of
keeping the personal condition or even in general the
social status, on its best if possible.

To generalize, one could say, part of the AI is
continuously checking the important environmental
variables which have direct, or close, but indirect
effects on it, and reacting accordingly to this.

This reaction will depend on the "personality". A set
of variables and constants which attempt to define how
the character would act in any kind of situation. 
The idea of a certain scoring system comes in handy
here for deciding.

I think this is as general as it gets. feedbacks?

How to link pieces of knowledge? now that is an
interesting thing to think about. In my opinion,
again, knowledge pieces should be identical in
structural definition. 
Shortly, they should be composed of a usage and a
condition, which may interact with the AI. 

Any other thoughts?

finding the most efficient {means} of reaching a goal
gains importance here.

Example: 
a horseshoe which is being made. It is in very hot
condition. The goal is already set. The AI needs to
hold it and change its place. The personality will
most likely try to avoid long term (in here, quite
short) contact through limbs with the hot condition
material, because it is in the harming subset. Seek
improvement of the condition by searching for
efficient objects to add into the "using" pool other
than its limbs. The AI will not be able to make a
machine to move the material perhaps, but a very
simple tool in the vincinity could be used to reach
the goal.

Wow. Even this simple action looks scary. A tool using
computer...

Another relaxing example:
Goal: reach from one destination to another.
AI should be able to check, if there is a tool (a
horse, a train) in the vincinity which it can use in
order to reach to the destination more effectively. 
----The more permanent the solution, the better.----

anyway sorry I got carried away...

EPITAPH-VII


__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail
http://personal.mail.yahoo.com/

Re: [Arianne-devel] About AI and object structures/by EPITHET-VII

[Peter S. (K. <pet...@ad...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wednesday 27 June 2001 20:53, you wrote:
> > Introduction:
> > -------------
> >
> > Throughout my experience, I realized one thing. An AI
> > should be able to set solid goals.
>
> Sure.
Throughout my literature, I know that goals shouldn't be known always, for 
example : you might be unable to specify the goal state (unknown+dynamic 
environment) but you can give most of the time a score to the present state 
of the algorithm/parameters. A search operator can try to maximize this 
score, thus performing potential usefull actions without (ever) reaching the 
goal state.

Unsupervised learning and reinforcement learning are a very well known 
example of this.

>
> > But it must be able to retreat from its goals in order
> > to gain advantage or change goal easily, whichever is
> > more beneficial.
> >
> > For the AI to have a goal, one needs a strictly
> > defined environment to which the AI will respond to.

- From your words i assume that you want to use a world representation based on 
(first order predicate) logic.  Which uses backtracking as you describe (when 
topdown SLD resolution is used for example)  But even in logic, people have 
found ways to handle incomplete knowledge, or even possibly false statements 
(enter Bayesian).

>
> Yes. But That AI makes plans to achieve complex task is just a good start
> point for me...
>
> > Now coming down to the object structure.. this is what
> > the AI is supposed respond to. If there will be a
> > structure, people who made the structure must be in
> > close contact with the people who create the
> > environmental- responsive AI.
> >
> > One could make a program over-intelligent, using great
> > amounts of energy and imagination.
>
> Well intelligence is related to the amount of information that the app

Uh ! rather the inverse is true ! Intelligence emerges from two main concepts 
: ability to generalise and ability to do that fast (search fast).  So keep 
your information limited(->faster) and fuzy (introduce 
variables/unknowns/don't cares)(->beter generalisation)
To much information will lead to overfitting when learning.

> has... We have to create that information from Perceptions and Actions...
> and NPC has to interchage that information so taht they can be as a single
> group. Uli has great ideas about this.
>
> > One could also make the program repetitive and stupid,
> > wanting to reach a simple goal.
>
> The key is that solving goals should add more goals to the NPC.

yep, this is topdown

of course you can only apply topdown proof procedures when the goalstate is 
known.  That's in my opinion a quite drawback of logic symbol manipulation 
systems, they are goal driven, so the/a goal must be presented by the 
environment/supervisor. that can be a hard thing to do and is also why a lot 
of these systems fail in natural environments, the environment demands that a 
certain goal is solved(cross the water without getting wet) but how do you 
tell this unexpected goal to the system ?
Very interesting to think about imho.

greetz, K.


>
> > These are just the ideas. Now I am thinking on the
> > AI 1.0  a program I started some time ago. No matter
> > how big the object structure is, it will try to
> > achieve the goal of
> > "changing the state or states of one or more objects
> > relative to the other objects"
> > sum(d.final)- sum(d.initial)= d/change= goal
> > I think this explains it all.
>
> Well. :-)
> I think that I need a bit more detailed explanation... :-)
>
> > Example: say the villain is supposed to assasinate
> > someone.
> > The goal is: villainhood- preserve your health
> > condition on the highest+ change the condition of your
> > subject to dead. How the villain does this will depend
> > on many many more environmental variables though, but
> > you get the idea.
> >
> > A learning program is different though... that I have
> > to consider later. Hope I am posting to the right
> > place.
>
> Yes, right place...
> Learning is easy, it is just to randomly choose between a random way of
> doing a goal or using a previous one...
>
> Regards,
>
>                   Miguel Angel Blanch Lardin
>
>              -- http://www.arianne.cx -- Arianne --
> The free open source massively multiplayer online role playing game
>
> nuclear cia fbi spy password code encrypt president bomb iran irak korea
> cuba Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola
> dna
>
>                     -- Echelon must die --
>
> _______________________________________________
> Arianne-devel mailing list
> Ari...@li...
> http://lists.sourceforge.net/lists/listinfo/arianne-devel

- -- 
- -------------------------------------------------
   Visit the 3Dsia Project http://www.3Dsia.org
 A Virtual Reality Environment without limits
           - now using RMC and ODB -
http://morgner.com/odb/   http://morgner.com/rmc/
- -------------------------------------------------
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7OyYFKfGwAkGiwccRAoIiAJ0U2OCVHhK4rrTdy4+Ive5f3slApACfVwW7
Y2EkTOzCGHSHydLJBjpcLtY=
=Z2EL
-----END PGP SIGNATURE-----

Re: [Arianne-devel] About AI and object structures/by EPITHET-VII

[<x51...@fe...>]

> Introduction:
> -------------
> 
> Throughout my experience, I realized one thing. An AI
> should be able to set solid goals. 

Sure.

> But it must be able to retreat from its goals in order
> to gain advantage or change goal easily, whichever is
> more beneficial. 
> 
> For the AI to have a goal, one needs a strictly
> defined environment to which the AI will respond to. 

Yes. But That AI makes plans to achieve complex task is just a good start point 
for me...
 
> Now coming down to the object structure.. this is what
> the AI is supposed respond to. If there will be a
> structure, people who made the structure must be in
> close contact with the people who create the
> environmental- responsive AI.
> 
> One could make a program over-intelligent, using great
> amounts of energy and imagination. 

Well intelligence is related to the amount of information that the app has...
We have to create that information from Perceptions and Actions... and NPC has 
to interchage that information so taht they can be as a single group.
Uli has great ideas about this.
 
> One could also make the program repetitive and stupid,
> wanting to reach a simple goal.

The key is that solving goals should add more goals to the NPC.

> These are just the ideas. Now I am thinking on the 
> AI 1.0  a program I started some time ago. No matter
> how big the object structure is, it will try to
> achieve the goal of 
> "changing the state or states of one or more objects
> relative to the other objects"
> sum(d.final)- sum(d.initial)= d/change= goal
> I think this explains it all. 

Well. :-) 
I think that I need a bit more detailed explanation... :-)

> Example: say the villain is supposed to assasinate
> someone. 
> The goal is: villainhood- preserve your health
> condition on the highest+ change the condition of your
> subject to dead. How the villain does this will depend
> on many many more environmental variables though, but
> you get the idea.
> 
> A learning program is different though... that I have
> to consider later. Hope I am posting to the right
> place.

Yes, right place...
Learning is easy, it is just to randomly choose between a random way of doing a 
goal or using a previous one...

Regards,

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] About AI and object structures/by EPITHET-VII

[ONUR T. <hal...@ya...>]

Introduction:
-------------

Throughout my experience, I realized one thing. An AI
should be able to set solid goals. 

But it must be able to retreat from its goals in order
to gain advantage or change goal easily, whichever is
more beneficial. 

For the AI to have a goal, one needs a strictly
defined environment to which the AI will respond to. 

Now coming down to the object structure.. this is what
the AI is supposed respond to. If there will be a
structure, people who made the structure must be in
close contact with the people who create the
environmental- responsive AI.

One could make a program over-intelligent, using great
amounts of energy and imagination. 

One could also make the program repetitive and stupid,
wanting to reach a simple goal.

These are just the ideas. Now I am thinking on the 
AI 1.0  a program I started some time ago. No matter
how big the object structure is, it will try to
achieve the goal of 
"changing the state or states of one or more objects
relative to the other objects"
sum(d.final)- sum(d.initial)= d/change= goal
I think this explains it all. 

Example: say the villain is supposed to assasinate
someone. 
The goal is: villainhood- preserve your health
condition on the highest+ change the condition of your
subject to dead. How the villain does this will depend
on many many more environmental variables though, but
you get the idea.

A learning program is different though... that I have
to consider later. Hope I am posting to the right
place.


__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail
http://personal.mail.yahoo.com/

Re: [Arianne-devel] [OT] test

[<x51...@fe...>]

Mensaje citado por: masahiro minami <el...@aa...>:

> Hi,
> 
> Sorry, this is a test, for i'm not getting any mail....

It works...
None use it :(
That is because I don't want to split the ML... :-)


                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] [OT] test

[masahiro m. <el...@aa...>]

Hi,

Sorry, this is a test, for i'm not getting any mail....

Regards,
-- 
masahiro minami <el...@aa...>

PowerBook G4/400
PowerMac G3/233
MacOS 9.1--MacOS X Public Beta---LinuxPPC
(Eudora 4.3.2J)
exception handling : www2.age.ne.jp/~except

Re: [Arianne-devel] Re: [Arianne-general] mysql++ on Linuxppc

[<x51...@fe...>]

Mensaje citado por: Ulrich Eckhardt <Doo...@kn...>:

> As I also had to be reminded, this rather belongs to the
> devel-mailinglist
> 
> 
> > account-manager.cpp:591
>  try {
>      db = new DataBase(1, database,"localhost",user, pass);
>  }
>  catch(...)
>  {
>      cerr << "Error: cought exception\n";
>      exit(1);
>  }
> 
> you could try adding more catch-clauses for std::exception and the
> various 
> SQL++-exceptions. That would at least give a hint on what is happening
> here.

Minami has a ugly problem  with MySQL++.
MySQL is running ok, because he can perfectly use it...

MySQL++ is not as good as we though on a first time I think...

Minami is working on porting it to bare MySQL, so that Arianne can run without 
MySQL++.



                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

[Arianne-devel] Re: [Arianne-general] mysql++ on Linuxppc

[Ulrich E. <Doo...@kn...>]

As I also had to be reminded, this rather belongs to the devel-mailinglis=
t


> account-manager.cpp:591
 try {
     db =3D new DataBase(1, database,"localhost",user, pass);
 }
 catch(...)
 {
     cerr << "Error: cought exception\n";
     exit(1);
 }

you could try adding more catch-clauses for std::exception and the variou=
s=20
SQL++-exceptions. That would at least give a hint on what is happening he=
re.

uli

Re: [Arianne-devel] [CODE] Approach

[Brian R. <row...@ha...>]

On 20 Jun 2001 12:53:17 +0200, x51...@fe... wrote:
> 
> > Wait for the requirements to be agreed before coding new 
> > areas.
> > 
> > An example:
> > coding is ongoing for the Renderer,and various other areas.
> > 
> > Why is waiting a short while before coding the Perception 
> > systems, or the Weather Model, when the discussions on 
> > how they work are still ongoing, such a difficult thing to do?
> 
> Yes, I must to agree with Quisar here...
> The whole system can't be waiting on a decision to be taken on a small area like 
>  perception is. IMO is better to discuss fast and then change code if we need 
> new requirements.

I agree with this idea... talk fast, change fast.  We can depend on cvs
to allow us to reverse bad changes.

But on _design_ changes or additions I think that we should be more
formal.

-- 
"Teach a man to make fire, and he will be warm for a day
Set a man on fire, and he will be warm for the rest of his life."
        - John Hrastar

Re: [Arianne-devel] [CODE] Approach

[<x51...@fe...>]

> Wait for the requirements to be agreed before coding new 
> areas.
> 
> An example:
> coding is ongoing for the Renderer,and various other areas.
> 
> Why is waiting a short while before coding the Perception 
> systems, or the Weather Model, when the discussions on 
> how they work are still ongoing, such a difficult thing to do?

Yes, I must to agree with Quisar here...
The whole system can't be waiting on a decision to be taken on a small area like 
 perception is. IMO is better to discuss fast and then change code if we need 
new requirements.

> > > There is, for the moment, very few people comitting code, and
> > > I really do not want this number to decrease one more time
> > > because of the rigidity of your system.
> > 
> > Yes, you are right.
> > But this method can save us a few emails of why this has been
> > coded or why code is not stable and so...
> 
> Having  a set of documents that lets newcomers know whats 
> going on, and how the various bits fit together, both as 
> interfacing modules, and as tasks in the overall planned 
> sequence of events, can only help make it easier for new 
> coders to join.
> 
> Making it easier for coders to slot in means an *in*creasing 
> number of coders.

Yeah, sound perfect, but: 
a) someone has to write the papers.
b) someone has to worry to update papers.

Really, that is pretty hard...
And we shouldn't add more burocrazy to the project, so that changing a line of 
code will mean to fill 50 forms and update 30 documents. 
But the idea take as a light-simple process could be good. 

                  Miguel Angel Blanch Lardin
 
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna 
 
                    -- Echelon must die --

Re: [Arianne-devel] [CODE] Approach

[Euan M. <lu...@us...>]

On 19 Jun 2001, at 15:35, x51...@fe... wrote:
> Mensaje citado por: Benjamin `Quisar' Lerman
> >  Euan Mee a =E9crit=A0: 
> > > What do you all think?
> > Sorry to be rude: BullShit.
> > I'm not gonna wait 3 weeks of dicussion before writing a line
> > of code, correcting a bug, changing some kind of
> > implementation that is not optimized or things like that...
>
> k, on that you are right. But we can decrease the timelines. And
> discuss on ML the functionality requiered.

I don't suggest at any time that there are 3 week coding 
holidays.

What I do say is: in a given functional area, we must agree 
what is to be coded, and how, before coding.

Why write code for something that's not wanted or needed, or 
which ignores the decisions on RP functionality?

Code all you like in the areas which are agreed on.

Wait for the requirements to be agreed before coding new 
areas.

An example:
coding is ongoing for the Renderer,and various other areas.

Why is waiting a short while before coding the Perception 
systems, or the Weather Model, when the discussions on 
how they work are still ongoing, such a difficult thing to do?

> > There is, for the moment, very few people comitting code, and
> > I really do not want this number to decrease one more time
> > because of the rigidity of your system.
> 
> Yes, you are right.
> But this method can save us a few emails of why this has been
> coded or why code is not stable and so...

Having  a set of documents that lets newcomers know whats 
going on, and how the various bits fit together, both as 
interfacing modules, and as tasks in the overall planned 
sequence of events, can only help make it easier for new 
coders to join.

Making it easier for coders to slot in means an *in*creasing 
number of coders.
 
> Anyway obvious things should be done in a automatic thing. To fix
> a bug is automatic, to design Object System is not... most of the
> code can be done without delay, the point is what to do, really
> howto do it is not so important.

Exactly.
Cheers,
     Euan
xl...@us...

'I would live all my life in nonchalance and insouciance,
Were it not for making a living, which is rather a nouciance'
 - Ogden Nash

Re: [Arianne-devel] [CODE] Approach

[<x51...@fe...>]

Mensaje citado por: Benjamin `Quisar' Lerman <ari...@qu...=
>:

>  Euan Mee a =E9crit=A0:=20
> > What do you all think?
>=20
>  Sorry to be rude: BullShit.
>=20
>  I'm not gonna wait 3 weeks of dicussion before writing a line of code,
> correcting a bug, changing some kind of implementation that is not
> optimized or things like that...

k, on that you are right. But we can decrease the timelines.
And discuss on ML the functionality requiered.

>  There is, for the moment, very few people comitting code, and I really
> do not want this number to decrease one more time because of the
> rigidity of your system.

Yes, you are right.
But this method can save us a few emails of why this has been coded or wh=
y code=20
is not stable and so...


Anyway obvious things should be done in a automatic thing. To fix a bug i=
s=20
automatic, to design Object System is not... most of the code can be done=
=20
without delay, the point is what to do, really howto do it is not so impo=
rtant.


                  Miguel Angel Blanch Lardin
=20
             -- http://www.arianne.cx -- Arianne --
The free open source massively multiplayer online role playing game

nuclear cia fbi spy password code encrypt president bomb iran irak korea =
cuba
Ala yihad mosad kgb free freedom human rights yugoslavia kosovo ebola dna=
=20
=20
                    -- Echelon must die --