At this moment the Spring SVN at BerliOS is used to develop the engine, lobby and some other things.
AI development is left out at this moment.
As I set myself the goal to streamline Spring development (across the full scope of the project) this can be a good thing to take a look at.
I would like to propose the inclusion of AI development in the Spring SVN.
Now I know the AAI and KAI AI's don't have there source included.
Why is this exactly? And could it be an idea to change this and move AI development inside the Spring SVN?
The source of the problem. :-)
Moderators: hoijui, Moderators
KAI is closed source and shall remain that way at least until the final release. No point developing things everyone will copy and give ill-credit for. How many AIs have kept the single text message line i asked for them to keep in order to use my metal class?
Plus getting involved with releases and getting delayed/delaying concurrent projects due to stability is also bad.
Having full access to the spring source and features would benefit KAI immensely (we are coding a lot of spring again, like pathfinding, ballistics, slopemaps etc etc), but that wouldnt be achieved by just including it in the SVN (it would have to be an integral part of spring), thus i see no real point to that , thanks anyways!
Plus getting involved with releases and getting delayed/delaying concurrent projects due to stability is also bad.
Having full access to the spring source and features would benefit KAI immensely (we are coding a lot of spring again, like pathfinding, ballistics, slopemaps etc etc), but that wouldnt be achieved by just including it in the SVN (it would have to be an integral part of spring), thus i see no real point to that , thanks anyways!
-
Lindir The Green
- Posts: 815
- Joined: 04 May 2005, 15:09
-
Tim Blokdijk
- Posts: 1242
- Joined: 29 May 2005, 11:18
It's getting late here so unfortunately I can't give a full response but thanks for the reply.krogothe wrote:KAI is closed source and shall remain that way at least until the final release. No point developing things everyone will copy and give ill-credit for. How many AIs have kept the single text message line i asked for them to keep in order to use my metal class?
Plus getting involved with releases and getting delayed/delaying concurrent projects due to stability is also bad.
Having full access to the spring source and features would benefit KAI immensely (we are coding a lot of spring again, like pathfinding, ballistics, slopemaps etc etc), but that wouldnt be achieved by just including it in the SVN (it would have to be an integral part of spring), thus i see no real point to that , thanks anyways!
I will pick this up tomorrow.
-
Tim Blokdijk
- Posts: 1242
- Joined: 29 May 2005, 11:18
Alright, the sun is up again (for some time) around here. 
Well, the problem with AI's is that it's cool to write a better AI then the other guy.
There is a form of competition that you don't get with general engine development.
But more cooperation would also be beneficial.
I'm really interested in exploring ways to overcome the disadvantages.
I would also like to hear Submarine's view on the topic.
AF, as you share your AI source (and already have the code in the Spring svn) would you like the commit access to develop it there?
You already found the mailinglist so it would make it easier to coordinate development with the rest of the developers.
Also in general (that is for everybody interested in AI development) it's fine to develop an AI in svn but there are some things that are important.
You first need to show that you can write an AI to get a place in svn.
The BerliOS admins (Jelmer, Tobi and Chris although he is not an active dev at this moment) decide how the svn is organized.
And there will be some limits on where AI developers can commit.
If you also like to commit to the Spring engine you have to show you can write clean code for it.
Expanding the use of Springs svn is a new idea that still needs to settle in and how things get organized will evolve in the coming months.
So if things can be organized more effectively, talk about it.
Well, the problem with AI's is that it's cool to write a better AI then the other guy.
There is a form of competition that you don't get with general engine development.
But more cooperation would also be beneficial.
I'm really interested in exploring ways to overcome the disadvantages.
I would also like to hear Submarine's view on the topic.
AF, as you share your AI source (and already have the code in the Spring svn) would you like the commit access to develop it there?
You already found the mailinglist so it would make it easier to coordinate development with the rest of the developers.
Also in general (that is for everybody interested in AI development) it's fine to develop an AI in svn but there are some things that are important.
You first need to show that you can write an AI to get a place in svn.
The BerliOS admins (Jelmer, Tobi and Chris although he is not an active dev at this moment) decide how the svn is organized.
And there will be some limits on where AI developers can commit.
If you also like to commit to the Spring engine you have to show you can write clean code for it.
Expanding the use of Springs svn is a new idea that still needs to settle in and how things get organized will evolve in the coming months.
So if things can be organized more effectively, talk about it.
-
Tim Blokdijk
- Posts: 1242
- Joined: 29 May 2005, 11:18
When you feel the need.
There are some guides on how to use svn.
You have the book http://svnbook.red-bean.com/en/1.1/ch01.html
And the chapter you would like to read http://svnbook.red-bean.com/en/1.1/ch03s05.html
And I found some some examples about good svn use here http://www.t2-project.org/svn-commit-tutorial.html
But it's all a bit Linux specific.
Maybe we could write something about using svn on windows, linux and mac at some time.
Anyway there are ways to do stupid things but you will manage fine.
Try some and if you can't figure it out ask on the mailinglist.
There are some guides on how to use svn.
You have the book http://svnbook.red-bean.com/en/1.1/ch01.html
And the chapter you would like to read http://svnbook.red-bean.com/en/1.1/ch03s05.html
And I found some some examples about good svn use here http://www.t2-project.org/svn-commit-tutorial.html
But it's all a bit Linux specific.
Maybe we could write something about using svn on windows, linux and mac at some time.
Anyway there are ways to do stupid things but you will manage fine.
Try some and if you can't figure it out ask on the mailinglist.
well the main reason why aai is closed source since version 0.3 is actually alantai (i dont want to start another flame war now, but i think af knows about it) - i guess most of you noticed that there had been several "fights" related to different topics like design of an ai, coding achievments etc. (i could go further into details if alantai wants me to do so, but my command of the english language is propably not good enough to put things into diplomatic phrases *g*) perhaps its time to settle this conflict.Tim Blokdijk wrote: I would also like to hear Submarine's view on the topic.
however there are quite a few ppl who got recent aai source (krogothe, nicke, accid_uk, tobi, etc) and krog and i have regular talks about our ais
Last edited by submarine on 02 Jul 2006, 14:48, edited 1 time in total.
-
unpossible
- Posts: 871
- Joined: 10 May 2005, 19:24
-
SwiftSpear
- Classic Community Lead
- Posts: 7287
- Joined: 12 Aug 2005, 09:29
Me?!
I asked actually asked for AAI source to be put up after the first closed source AAI and you promised you would, and I was very dissapointed when you turned round and said it was now closed source. So please clarify for me why you felt the need to go clsoed source because of me?
In all of this you forget that we can take code form eachother but if the author suddenly revokes that right then it has to be totally rewritten from scratch or unintegrated.
I asked actually asked for AAI source to be put up after the first closed source AAI and you promised you would, and I was very dissapointed when you turned round and said it was now closed source. So please clarify for me why you felt the need to go clsoed source because of me?
In all of this you forget that we can take code form eachother but if the author suddenly revokes that right then it has to be totally rewritten from scratch or unintegrated.
just a short answer, im suffering from a bad headache today *argh*
you are right, i said i would post the source later... but thats not what i was refering to - i was rather thinking of your behaviour. you probably spend more time on your ai and ais in general than any of us but i cant get rid of the feeling, that there is a huge gap between your annoucements and your achievements. you may have hundreds of reasons to refute that - but remember it's just a feeling (but i know im not the only one who thinks that way)
you are right, i said i would post the source later... but thats not what i was refering to - i was rather thinking of your behaviour. you probably spend more time on your ai and ais in general than any of us but i cant get rid of the feeling, that there is a huge gap between your annoucements and your achievements. you may have hundreds of reasons to refute that - but remember it's just a feeling (but i know im not the only one who thinks that way)
no i wasnt referring to epic - hmm its difficult to explain:
you tend to announce with each ntai release that it will now be the "ultimate uber killer ai", totally bug free and uncrashable
in the end i was always dissapointed by lots of (crash)bugs, the poor speed (though it seems you optimized ntai quite a lot during the last versions), the poor overall game performance (last version i tried was xe8.5 though). its not that the other ais are so much better - i just have the feeling that the other ai devs produce less hot air and boast much less
you tend to announce with each ntai release that it will now be the "ultimate uber killer ai", totally bug free and uncrashable
in the end i was always dissapointed by lots of (crash)bugs, the poor speed (though it seems you optimized ntai quite a lot during the last versions), the poor overall game performance (last version i tried was xe8.5 though). its not that the other ais are so much better - i just have the feeling that the other ai devs produce less hot air and boast much less
Well when i release them they do seem uncrashable and bug free, as I've found the bugs and quashed them. Only past experience has meant that earlier versions of NTai had bug sin them because I fiddled something just before my Internet time was about to begin and didn't have time to test, which has changed now, my builds are more reliable than they used to be.
I'm not t he only person who 'boasts' about AI's though I don't intend to boast unless I'm writing a release post or a news item, or the many occasions where I'm trying to persuade people to upgrade their version of NTai (curse you NTai XE7.5 being bundled with spring)
Of course nowadays I get told of for simply mentioning NTai, e.g. I made a comment in the KAI thread in reply to someone who said something about NTai, it was a boast, yet people snapped at me for advertising and boasting. And in #main when Noize asked me a load of questions, he then went on at me for boasting about NTai after i answered his questions. I cant help but mention the word NTai if I'm responding to an NTai related question....
How that relates to AAI being closed source exactly isn't 100% clear though.
btw Krogothe, I didn't remove that line of t he metal algorithm, I simply changed ti to use the logging class. You can see it in all the log files NTai generates and I'm sure there's a message in the console saying credit to cain/Krogothe/Veylon. OTAI still has the line, and AAI still has it as far as I'm aware, just that our AI's might output so much to the console we have to trim it a little to fit it on.
I'm not t he only person who 'boasts' about AI's though I don't intend to boast unless I'm writing a release post or a news item, or the many occasions where I'm trying to persuade people to upgrade their version of NTai (curse you NTai XE7.5 being bundled with spring)
Of course nowadays I get told of for simply mentioning NTai, e.g. I made a comment in the KAI thread in reply to someone who said something about NTai, it was a boast, yet people snapped at me for advertising and boasting. And in #main when Noize asked me a load of questions, he then went on at me for boasting about NTai after i answered his questions. I cant help but mention the word NTai if I'm responding to an NTai related question....
How that relates to AAI being closed source exactly isn't 100% clear though.
You do seem to be putting this diplomatically, but that's what i missed a lot about the past, people used to talk with me about theory and ideas and suggestions, but people don't talk about that them anymore, it's the little things like "units should guard commanders" or "the buildtree would be better with this" or "that sounds like a good idea". This whole atmosphere seems to have killed it off. I have so many more ideas I'd wanna talk about but I don't, I could have started loads of new threads in this forum about all these new ideas, but the whole KAI super AI this and that, and the "that's a good idea, when can it be done" sort of reaction doesn't serve those ideas any dignity.however there are quite a few ppl who got recent aai source (krogothe, nicke, accid_uk, tobi, etc) and krog and i have regular talks about our ais
btw Krogothe, I didn't remove that line of t he metal algorithm, I simply changed ti to use the logging class. You can see it in all the log files NTai generates and I'm sure there's a message in the console saying credit to cain/Krogothe/Veylon. OTAI still has the line, and AAI still has it as far as I'm aware, just that our AI's might output so much to the console we have to trim it a little to fit it on.
-
unpossible
- Posts: 871
- Joined: 10 May 2005, 19:24
I choose to not pay much attention to new ideas due to the likes of kelson and his metal class dreams, or the voronoi diagrams that (forgive me if im wrong) never saw use. At least Gabba is a friendly, helpful guy, while kelsons been an idiot who never did anything remotely useful to do with metal classes. He bragged about his uber groupais no ones seen yet. People like him make me stubborn, especially when the stubborn way has worked so well so far.
Lastly, i think if someone has a good idea thats worth putting time into they should be motivated and secure enough to let it be known even with the inevitable criticism.Meeting fierce resistance is a part of any ideological revolution, so live with it. A working example does wonders to convince people instead of all talk too
Lastly, i think if someone has a good idea thats worth putting time into they should be motivated and secure enough to let it be known even with the inevitable criticism.Meeting fierce resistance is a part of any ideological revolution, so live with it. A working example does wonders to convince people instead of all talk too
Ideas are important and need nurturing and one very valuable resource is feedback and refinement of those ideas.
That has not happened here in a very long time, and self refinement however effective it is can only go so far. I've turned to a new concept of merging them together amassing lots of new ideas, rather than refining the old ones.
Without the refinement process that sued to occur and the old atmosphere stbbornness entrenches, and it is stubbornness that killed off QAI, SAI, ZcAIn, MyAI, TAI, and 5 matrix AI, and if we dont hear anything from veylon, OTAI.
Stubbornness is what drove kelson out. Lack of knowledge however si what killed off voronoi diagrams. Many people have come here who kenw how, and I was happy to see that, then greatly dissapointed when they ran away to other communities without a word of goodbye. The current state only serves to promote the 3 major AI's into the colelctive monopoly they have become, trying to outdo eachother.
That has not happened here in a very long time, and self refinement however effective it is can only go so far. I've turned to a new concept of merging them together amassing lots of new ideas, rather than refining the old ones.
Without the refinement process that sued to occur and the old atmosphere stbbornness entrenches, and it is stubbornness that killed off QAI, SAI, ZcAIn, MyAI, TAI, and 5 matrix AI, and if we dont hear anything from veylon, OTAI.
Stubbornness is what drove kelson out. Lack of knowledge however si what killed off voronoi diagrams. Many people have come here who kenw how, and I was happy to see that, then greatly dissapointed when they ran away to other communities without a word of goodbye. The current state only serves to promote the 3 major AI's into the colelctive monopoly they have become, trying to outdo eachother.