Well, Tsumari, I am not trying to be elitist, I am being selective by letting those in who i think can help and pushing those who would be a negative source be pushed out.
I have no intention fo turning spring into the sort fo community where everyone keeps everything private.
However AI and maps are not the same, nor are client ports and AI's the same, I am in a position where I have rival projects and thus have competition and in light of that have a vested interest in the guarding of what ideas and concepts I or anyone on darkstars brings with them.
This is an AI, it wont require a forking of the spring code as long as Jouninkomiko provides an AI interface and doesnt leave us all havign to code ti ourselves, maybe then spring might indeed fork if we didnt all agree and no SY intervened but I find that unlikely.
As for private entities falling apart at any time, I have every intention of seing this through my ideas and thoughts till I see them infront of me playing em at spring, and playing it well at that, and wether triaxx doesnt do it cus he thinsk his coding skills arent what they should be or you get bored or Jouninkomikos AI suddenly appears oneday when we werent expecting it, I'm still going todo it. I've committed myself to it.....
Afterrall Spring si the primary project, any AI project is a secondary even tertiary project in relation to spring.
Global AI Interface
Moderators: hoijui, Moderators
Good reasoning you have here, tsumari.Tsumari wrote:I will not join a closed up project for Spring. If I contribute to Spring it will be in these forums this year, and next year I may take some tasks to code in the Summer.
...
I will contribute any ideas i have to the open, main ai-subproject, and quite possibly some code (ideas) aswell.
I'd also be happy to help write documentation to explain the concepts used in the AI engine to 'ordinary programmers' who don't yet have a background in gameAI coding.
As for me joining alantai's team to read up on their stuff, i know that's a bit of a paradoxal move in light of his nondisclosure wishes, which i've accepted.
but just as programs can be reverse engineered and their principles and operation copied, so can ideas and concepts for engines be re-used without that affecting highlevel endperformance of competing products.
Sorry for the load of philosophy here, but i don't want to start nervousness in the community at any point.
From both sides of the issue I can understand the problems, however, we want to remain closed source until it's complete, because otherwise someone is going to yell at us about it being buggy if they attempt to implement it before we finish. That's all there is to it. At least when it comes to mine.
I have no problem with closed source, in fact when I was a member of the SR team I was a huge advocate of keeping everything closed up. If you guys have the programming skills to accomplish what you set out then you're not hurting Spring or your own project. However, if you fall short or head down a path that won't work you will be pulling resources away from what could be some nice Spring development. But that's not my primary point.Triaxx2 wrote:From both sides of the issue I can understand the problems, however, we want to remain closed source until it's complete, because otherwise someone is going to yell at us about it being buggy if they attempt to implement it before we finish. That's all there is to it. At least when it comes to mine.
Spring, to have gotten this far, must have really good teamwork as well as programming skills. Using what they've built as a team to further progress in other areas will be more efficient, I believe, than cobbling together a new group which may or may not gel. I have also seen the attitude and motion of Alantia before in myself and others, and if he's going to think the way he does he needs to have a hell of a resume behind him, or someone with a hell of a resume to write the code he is going to be asking for.
Tsumari
And I see your point and I say you have not quite realised what I have said. I will start leaking thigns slowly once we've gotten far ahead or fi it's obvious your makign very similair paths here.
This is not a team locked under a great iron bolt on a huge safe door. Anyone can join the team as long as I see they can be trusted to follow the rule of no leak, I dont midn fi they've never heard of spring and never intend on playing it if they just want to deal with an RTS AI. I say what I say because this is not a TAI & 5 matrix AI dev team totally deedicated to what has already been posted and creatign code and using strict guidelines. I want people who have access to darkstars to be innovative, intuitive and to challenge whats been posted ro follow a completely different AI path if they wish, it's all research and being innovative and learning.
Fine we will need coders and I am fully aware of my own coding skills and where they are headed and that the people with access may grow over time as I want it to.
My main reasons for locking up access int he way I ahve doen si to keep hostile people out, people who care not for me and the others and wish mearly to learn what we have learnt and not contribute to anyone but themselves. I dont mind havign multiple AI paths actually being implemented and released from there if it means progress as logn as there's a friendly and contributive atmosphere about the whole thing.
And yes I udnerstand your concerns about it and know that what you have experienced yourself tsumari is what makes you think this isnt right. This is pretty much invite only not waltz in and ring up a storm.
As I have said I will post my ideas and develop on Darkstars and I will release things that I deem necessary or already caught upon by the community here. But I dont think this message board is a healthy place for ideas on AI, and I cant help feel that this may end up as another logo fiasco thread turned forum and project.
Tsumari, I have said I am committed to the project and I will not let anythign damage it, if the setup needs to change to continue it will change and I will nto let the whole AI thing collapse, I owe you all that much at least.
Fine, if you want a community project you can stay here at the Spring board and post AI ideas, or you can come over to Darkstars and read what we have to offer you and post your ideas there, sufficive to say the bulk of contributing idea makers concerning AI of the forums before spring where released have moved to darkstars or have a standing invitation. Dont elt the new people lag behind making ideas and paths we've already tread along in a place like this.
But sufficive to say, I want you there Tsumari, at least to see thigns and to advise us if you think whats happening has gone down the pan, at least to just give an opinion, you can be the lookout for us all if you think we'll cause damage just to safeguard. Just be aware that I trust you to keep the rule unbroken and that you wont dissapoint me. If you think a particular item is relevant and should be released then I'll take you seriously as any other person at darkstars whor equests releases. Please dotn think of me as the greedy librarian hoarding information to suit nobody but my own ends with no regard for the Spring project as a whole and the people here. I'm just doing what I think is right..
All it takes is a PM to join..........
This is not a team locked under a great iron bolt on a huge safe door. Anyone can join the team as long as I see they can be trusted to follow the rule of no leak, I dont midn fi they've never heard of spring and never intend on playing it if they just want to deal with an RTS AI. I say what I say because this is not a TAI & 5 matrix AI dev team totally deedicated to what has already been posted and creatign code and using strict guidelines. I want people who have access to darkstars to be innovative, intuitive and to challenge whats been posted ro follow a completely different AI path if they wish, it's all research and being innovative and learning.
Fine we will need coders and I am fully aware of my own coding skills and where they are headed and that the people with access may grow over time as I want it to.
My main reasons for locking up access int he way I ahve doen si to keep hostile people out, people who care not for me and the others and wish mearly to learn what we have learnt and not contribute to anyone but themselves. I dont mind havign multiple AI paths actually being implemented and released from there if it means progress as logn as there's a friendly and contributive atmosphere about the whole thing.
And yes I udnerstand your concerns about it and know that what you have experienced yourself tsumari is what makes you think this isnt right. This is pretty much invite only not waltz in and ring up a storm.
As I have said I will post my ideas and develop on Darkstars and I will release things that I deem necessary or already caught upon by the community here. But I dont think this message board is a healthy place for ideas on AI, and I cant help feel that this may end up as another logo fiasco thread turned forum and project.
Tsumari, I have said I am committed to the project and I will not let anythign damage it, if the setup needs to change to continue it will change and I will nto let the whole AI thing collapse, I owe you all that much at least.
Fine, if you want a community project you can stay here at the Spring board and post AI ideas, or you can come over to Darkstars and read what we have to offer you and post your ideas there, sufficive to say the bulk of contributing idea makers concerning AI of the forums before spring where released have moved to darkstars or have a standing invitation. Dont elt the new people lag behind making ideas and paths we've already tread along in a place like this.
But sufficive to say, I want you there Tsumari, at least to see thigns and to advise us if you think whats happening has gone down the pan, at least to just give an opinion, you can be the lookout for us all if you think we'll cause damage just to safeguard. Just be aware that I trust you to keep the rule unbroken and that you wont dissapoint me. If you think a particular item is relevant and should be released then I'll take you seriously as any other person at darkstars whor equests releases. Please dotn think of me as the greedy librarian hoarding information to suit nobody but my own ends with no regard for the Spring project as a whole and the people here. I'm just doing what I think is right..
All it takes is a PM to join..........
Ideas are meaningless unless they can be translated into algorithms to be implemented. The second stage is far, far more difficult than the first, which is why I've tended to suggest ideas leaning toward the simple.
If you want a super smart AI who understands the map you are going to either need to outright tell it about what the key features of a map are, or you're going to have to produce what will need to be highly skilled code to read a bunch of heightmaps, compare that to start positions, and try to make sense of coordinates with a mathematical function to determine what the key elements of a map are.
Your AI team, if it achieves real progress (and I wish it no ill will) is going to relatively quickly move away from the "bold overall idea" discussion and into things like, "How does an AI decide if it should build ships in a particular body of water? Can it tell if this body of water is taken by the enemy? Does it know when to attrit an area with artillery instead of attack it?" And those types of questions will lead to, "Who lost the damn struct pointer?" or "My function is somehow returning an invalid coordinate."
Now all this said, I am not a god coder of any sort. In fact, the only language I am really proficient in is C and maybe VB. I'll be doing C++ in the next year or so. I just want people to realize where the real effort is, and it won't necessarily be in writing up the document that basically boils down to, "This is how the AI will be smart." I worry because you are saying you want an unbeatable AI, and it won't happen. I can look at a map for 20 seconds and discern more detail than years of your team's coding will achieve.
Tsumari
If you want a super smart AI who understands the map you are going to either need to outright tell it about what the key features of a map are, or you're going to have to produce what will need to be highly skilled code to read a bunch of heightmaps, compare that to start positions, and try to make sense of coordinates with a mathematical function to determine what the key elements of a map are.
Your AI team, if it achieves real progress (and I wish it no ill will) is going to relatively quickly move away from the "bold overall idea" discussion and into things like, "How does an AI decide if it should build ships in a particular body of water? Can it tell if this body of water is taken by the enemy? Does it know when to attrit an area with artillery instead of attack it?" And those types of questions will lead to, "Who lost the damn struct pointer?" or "My function is somehow returning an invalid coordinate."
Now all this said, I am not a god coder of any sort. In fact, the only language I am really proficient in is C and maybe VB. I'll be doing C++ in the next year or so. I just want people to realize where the real effort is, and it won't necessarily be in writing up the document that basically boils down to, "This is how the AI will be smart." I worry because you are saying you want an unbeatable AI, and it won't happen. I can look at a map for 20 seconds and discern more detail than years of your team's coding will achieve.
Tsumari
It is jouninkomiko who wants an unbeatable AI, ti si I who wants an AI that follows the teachings of Sun Tzu.
And we have already dicerned solutions to those problems tsumari, some of which have been on this forum for months.
We have so far taken many problems which forum goers had thought would require complex algorithms and replaced them with simple dieas that can eb expanded into other areas of AI.
For example a threat matrix with grid squares. Thus we now ahve an efficient way of placing buildings, determining who si winning based on land control, and other solutions and ways of determining thigns that would otherwise requrie complex algorithms that are prooven to be simply unnecessary.
To be honest I see you seem to be causing more damage to the community than Darkstars and my attitude to leaks fo data.
Ideas are everything, without them we have nothign to implement, we are stupid, meaningless and devoid of any spiffyness or inspiration.
people = Ideas = theories = design&planning = implementation = end result
And I am aware of the implementation, I do know standard C++ & C & VB6 and know a variety fo thigns about other languages such as x86 assembly & Java & perl &python and a lot fo how the system itself manages code and memory, and I keep that in mind with ym ideas, I do nto continue wiht an diea if I see it is computationally impossible and look for a better way fo doing ti, it's innovation and resourcefullness.
Yet instead of doing my usual thign i am drawn into this thread defending my actions and trying to make you truly see.
Read every AI related thread in this forum and be sure to send the links to the threads to Jouninkomiko who ahsnt the time to look for and move them to this forum. They'll be in General Discussion, and I demand that you read each and every thread and each and every post in them before you come back.
This forum is being consumed by this. Go create a few threads with your ideas, and start the ball rolling here if you're intent on staying, dont get rid of what value we have left for AI in this community and alienate those of us who are already here before the next wave of people arrive to grace these forums.
And we have already dicerned solutions to those problems tsumari, some of which have been on this forum for months.
We have so far taken many problems which forum goers had thought would require complex algorithms and replaced them with simple dieas that can eb expanded into other areas of AI.
For example a threat matrix with grid squares. Thus we now ahve an efficient way of placing buildings, determining who si winning based on land control, and other solutions and ways of determining thigns that would otherwise requrie complex algorithms that are prooven to be simply unnecessary.
To be honest I see you seem to be causing more damage to the community than Darkstars and my attitude to leaks fo data.
Ideas are everything, without them we have nothign to implement, we are stupid, meaningless and devoid of any spiffyness or inspiration.
people = Ideas = theories = design&planning = implementation = end result
And I am aware of the implementation, I do know standard C++ & C & VB6 and know a variety fo thigns about other languages such as x86 assembly & Java & perl &python and a lot fo how the system itself manages code and memory, and I keep that in mind with ym ideas, I do nto continue wiht an diea if I see it is computationally impossible and look for a better way fo doing ti, it's innovation and resourcefullness.
Yet instead of doing my usual thign i am drawn into this thread defending my actions and trying to make you truly see.
Read every AI related thread in this forum and be sure to send the links to the threads to Jouninkomiko who ahsnt the time to look for and move them to this forum. They'll be in General Discussion, and I demand that you read each and every thread and each and every post in them before you come back.
This forum is being consumed by this. Go create a few threads with your ideas, and start the ball rolling here if you're intent on staying, dont get rid of what value we have left for AI in this community and alienate those of us who are already here before the next wave of people arrive to grace these forums.
Both of you should calm down (its just a comercial...)
But Alantai Firestar does have a point. You only joined in may. Even if you were here befor that... were u here from the begining? We've been working on thsi for some time and have plenty of ideas and a small willing group, and even if it turns out bad, we WIll make an AI.
I knwo your not against it. Just dont diss how were going about it or question our ability.
An translating ideas into algorithms is not as hard as it sounds. You just have to use a bit of logical thinking, and make it in a way a computer would understand. you hardly even need to know how to program. then once you know what you want it to do you either code it or give it to someone else to code, so realy... i dont see how you can say The second stage is far, far more difficult than the first. Cause it isn't. It just depends how u think.
aGorm
*some of that may be nonsence*
But Alantai Firestar does have a point. You only joined in may. Even if you were here befor that... were u here from the begining? We've been working on thsi for some time and have plenty of ideas and a small willing group, and even if it turns out bad, we WIll make an AI.
I knwo your not against it. Just dont diss how were going about it or question our ability.
An translating ideas into algorithms is not as hard as it sounds. You just have to use a bit of logical thinking, and make it in a way a computer would understand. you hardly even need to know how to program. then once you know what you want it to do you either code it or give it to someone else to code, so realy... i dont see how you can say The second stage is far, far more difficult than the first. Cause it isn't. It just depends how u think.
aGorm
*some of that may be nonsence*
Oh and Tsumari, try to break complex large thigns into smaller thigns. For exampel you said dealing with map features? Does the AI reallyhave to deal with ti as a whole, we can just program it to look for and recognise patterns on the map that are distinctive of canyons or passes or plains etc and if it cant do that then threat values will unveil the area in question as a strategic strongpoint.
We have hope, we have a vision. Join us, dont spoil it for us.
We have hope, we have a vision. Join us, dont spoil it for us.
-
jouninkomiko
- Posts: 436
- Joined: 26 Aug 2004, 08:11
ehm....Alantai Firestar wrote: For example a threat matrix with grid squares. Thus we now ahve an efficient way of placing buildings, determining who si winning based on land control, and other solutions and ways of determining thigns that would otherwise requrie complex algorithms that are prooven to be simply unnecessary.
hunters work by knowing their prey, and the routes to the resources the prey uses.
how does a threat matrix deal with predicting the paths an enemy uses, and could use?
note; i'm thinking about complicated maps here, not the simple ones we're seeing today in spring.
as you know, there is already a 'pathfinder' object in the spring code, so determining all possible efficient routes should be fairly easy too.
see my homepage for further details on my expectation levels for this subproject..
jouninkomiko, I'm not that much of a programer. I can probablie (if I found a book to refresh myself) get by in VB.
But once you know what he computer likes... it isn't so hard. I suppose being good at maths and knowing how to describe things in a mathamatical way helps a lot.
Maybe It's not the case
aGorm
But once you know what he computer likes... it isn't so hard. I suppose being good at maths and knowing how to describe things in a mathamatical way helps a lot.
Maybe It's not the case
aGorm
-
PauloMorfeo
- Posts: 2004
- Joined: 15 Dec 2004, 20:53
I just hope that the development of TA::Spring never gets dependent on closed development.
When, for example, an AI is beeing closedly developed, if the project fails for some reason (team members leaving, people loosing they're interest, etc, etc), most probably everything will be lost. Usually, all work is lost, literally, and if not, what's left of the work is most certainly not pickable by others.
However, if the project is not dependent on closed development parts, i don't see why there is a problem with that. Imagine that we didn't even knew about a certain project existance... We wouldn't have any problem with it getting lost, we wouldn't even know.
aGorm:
I bet you haven't started coding yet... When you do, you'll see how hard it is to cope with the limitation imposed by programming.
Anyway, i really hope you guys can develop a very good AI.
May the gods of open source be with you!
When, for example, an AI is beeing closedly developed, if the project fails for some reason (team members leaving, people loosing they're interest, etc, etc), most probably everything will be lost. Usually, all work is lost, literally, and if not, what's left of the work is most certainly not pickable by others.
However, if the project is not dependent on closed development parts, i don't see why there is a problem with that. Imagine that we didn't even knew about a certain project existance... We wouldn't have any problem with it getting lost, we wouldn't even know.
aGorm:
I bet you haven't started coding yet... When you do, you'll see how hard it is to cope with the limitation imposed by programming.
Anyway, i really hope you guys can develop a very good AI.
May the gods of open source be with you!
And just remember that this isnt a strict access thing. If you have ideas and show us thema nd are trustworthy then your allowed in. Just ask and you'll be elt in as logn as you dont break the no leak rule.
In the event of such a catastrophe darkstars would be opened up compeltely but I doubt such a catastrophe will ever happen nor will people allow Spring to rely upon the AI projects at darkstars to have a monopoly on skirmish.
In the event of such a catastrophe darkstars would be opened up compeltely but I doubt such a catastrophe will ever happen nor will people allow Spring to rely upon the AI projects at darkstars to have a monopoly on skirmish.
Oh dont wory I know how anyoying it can be when the computer dosen't understand the line:
Get the damed X co-ordinate of that bloddy shape and its write it to a sodding varible...
If you could code like that everyone would be coding.
I suppose this will mean ill have to put my money where my mouth is at some point though... dam.
aGorm
Get the damed X co-ordinate of that bloddy shape and its write it to a sodding varible...
If you could code like that everyone would be coding.
I suppose this will mean ill have to put my money where my mouth is at some point though... dam.
aGorm
Various responses to different things.aGorm wrote:Both of you should calm down (its just a comercial...)
An translating ideas into algorithms is not as hard as it sounds. You just have to use a bit of logical thinking, and make it in a way a computer would understand. you hardly even need to know how to program. then once you know what you want it to do you either code it or give it to someone else to code, so realy... i dont see how you can say The second stage is far, far more difficult than the first. Cause it isn't. It just depends how u think.
aGorm
*some of that may be nonsence*
If you don't know how to program you probably can't design proper alrgorithms. That's like saying you can get derivatives without knowing math. If you're not a programmer on a project such as an AI all you can do is make wishlists and playtest. A wishlist is not an algorithm. A wishlist is very easy to make.
There's not really any point in talking about who's been in this forum longer, since Spring just opened up recently. I wasn't going to BS about a project I knew relatively little about. I've been on TAU for what, 7 years? But none of that really impacts the discussion.
I'm not questioning anyone's abilities because I don't know what those abilities are. That's important to note because if that is misunderstood then it will just make people upset.
Alantia, about the canyon: Being able to recognize something like that was the difficulty I was talking about. If you get the AI to realize it has a chokehold you also have to figure out a way to make it decide if it is one that matters.
Telling me to go away because I supported the idea of open development is pretty ironic. I doubt I'm hurting any development - I think I'm being pretty civilized, I just happen to disagree with you.
The forums were wiped so I can't go back and read the archives. That's not something I can fix. I won't join your AI project because 1) you don't need me and 2) I won't acccept an NDA.
Tsumari
Well Tsumari, the forums where wiped and the only AI thread at the time was SSAI (please see archived data I posted in the SSAI tidbits thread.) It was after the forums where wiped that the TAI and 5 matrix threads where created aswella s their associated threads.
Also I think you have been ignorant of what abilities we have and should never ahve mentioend them int he first palce.
Also the problem with dientifyign a choke point is not to take control fot h choke point, rather a different cocnept is needed. Perhaps a way of destroying the choke point or a chance of deceit or manipulation on behalf fo the AI. Such as moving the choke point, creatign a minefield, obliteratign the choek point into a quagmire of craters, etc
Also I think you have been ignorant of what abilities we have and should never ahve mentioend them int he first palce.
Also the problem with dientifyign a choke point is not to take control fot h choke point, rather a different cocnept is needed. Perhaps a way of destroying the choke point or a chance of deceit or manipulation on behalf fo the AI. Such as moving the choke point, creatign a minefield, obliteratign the choek point into a quagmire of craters, etc
I've been working on the choke point problem since it was mentioned, and I have a partial solution to it at least. Any location that can be identified as a chokepoint, is considered a High threat, medium value area, and is defended accordingly. My AI uses mines offensively, rather than defensively, so mining the pass isn't going to help me much.