The Opensource paradox

Post by **jcnossen** » 09 Dec 2005, 02:10

Ok then I will release my AI for my hypothetical closed source RTS engine in the future, which happens to use the same interface

You see the problem here?, you can't release the same thing under two different licenses at a time...

Post by SJ » 09 Dec 2005, 20:03

But you are still using the .h files that I created (and you have modified) and released under the GPL. And the GPL only allows you to use GPLed files in a project if you release the whole project under GPL. Probably you could create some reverse enginered .h files I suppose (well not you since you have obviously seen the original files and thus cant do a clean reverse enginering but someone that has never looked on the spring source).

Triaxx2 · Post by **Triaxx2** » 09 Dec 2005, 20:22

Look at it this way. Without your .h files, my code still works, it just doesn't have the interface. If I create .h files of my own, and retarget my source to those files... Not to mention that to work with another engine, it'd more than likely have to be compiled to a different format.

AF · Post by AF » 09 Dec 2005, 20:38

hmm, more to the point not every AI has its latest version posted as source, and no AI has labelled itself closed source.

NTAI, OTAI, JCAI,ZcAIn Opensource

AAI unverified opensource post 0.22

Great list fo incmplete or planend AI's unknown

What point is ther ein putting NTAI source code up for other devs to see if they wont return the courtesy.

Post by **jcnossen** » 09 Dec 2005, 23:42

But you are still using the .h files that I created (and you have modified) and released under the GPL. And the GPL only allows you to use GPLed files in a project if you release the whole project under GPL. Probably you could create some reverse enginered .h files I suppose (well not you since you have obviously seen the original files and thus cant do a clean reverse enginering but someone that has never looked on the spring source).

Ok true... I will have to release at least the parts that bind my AI code to spring.
This is fair hypothetical though, I'm not working on AI anyway... but I just don't like these restrictions because I think they make the AI competition less fun. Any other AI competitions (such as all those things with virtual programmable robots in an arena) have closed source entries as well.

submarine · Post by **submarine** » 10 Dec 2005, 00:21

i dont think its necessary to upload new source with every bugfix.

anyway aai inspired many other people and aai code has been implemented in at least two other ais and there were two other people who told me they want to start dev. their own ai based on aai

krogothe · Post by **krogothe** » 10 Dec 2005, 00:40

Sub is right, his AI is a major contribution to my AI in terms of code, and even though i'll be rewriting that code on sunday, leaving no trace of AAI, its help in getting me started and teaching me about c++ cannot be denied!
I believe publishing code should be encouraged, not enforced... I posted my Metal Class, and ill be glad if it helps people (especially with the atrocious mex placing in most released AIs to date

). I wouldnt be glad at all if i was made to do it however...
Im sure sub will release the source to AAI once v0.3 is complete... As should I once my AI actually does something?

Nemo · Post by **Nemo** » 10 Dec 2005, 03:23

I don't think its a big deal...until we start doing some form of organized AI competitions. When that sort of thing starts up, I think its only fair that the AI makers in that are allowed to keep their source secret, at least for the duration of the contest.

That way, everyone can speculate about how AAI does that nifty trick it does, ect, and then when the tourney is over, you get to see what makes them tick.

I suppose this violates GPL, but if the source is eventually released (and is being withheld for something like a fun and bragging rights-only contest), is it that big of a deal?

I'm really looking forward to AI vs AI tournaments. A little competitive spirit never hurt anyone, and might spark some even more amazing AIs.

Neuralize · Post by **Neuralize** » 10 Dec 2005, 03:26

Personal Attack: I just think Alantai doesn't want anyone sneaking up on the utter "greatness" that is his AI.

Triaxx2 · Post by **Triaxx2** » 10 Dec 2005, 03:57

No personal attacks.

And if we're doing AI competitions, then the source should be restricted access at least for the duration. Perhaps requiring permission from two different people?

submarine · Post by **submarine** » 10 Dec 2005, 12:19

i will release source for all major releawses, i#m just too lazy to upload new source for every single bugfix...

Post by **Tim Blokdijk** » 10 Dec 2005, 15:02

Zaphod wrote:...
Ok true... I will have to release at least the parts that bind my AI code to spring.
This is fair hypothetical though, I'm not working on AI anyway... but I just don't like these restrictions because I think they make the AI competition less fun. Any other AI competitions (such as all those things with virtual programmable robots in an arena) have closed source entries as well.

It's hypothetical but still good to know the implications of the GPL on the Spring project.
I would like to keep AI code under the GPL.

For AI competitions you can keep you AI private for 1-2 months and then release it at a predefined date together with the other players. (and with the source)
The version released on that date can battle around in a competition.
At the same time you can study the code made by the others and use it for the next competition 1-2 months later.
Nobody violates the GPL and you still have healthy competition.

AF · Post by AF » 10 Dec 2005, 17:16

aha, Neuralize, you've described how I developed NTAI before Istarted bundling th source with every release and doing quick releases that ahdnt bene fulyl tested.

Sub, your code is there yes, btu a bugfix is a bugfix, it may show others who used your code that they have a bug to fix, ro ti may show them bugs in their own code they hadnt realized. In developing NTAI I see a bug, then i look for similair instances in which ti might occur, and even though they might not eb a big problem that I've even experienced at all the new bugs I find are fixed. The same was true of watching the bugfixes Zaphod made to JCAI.

You've put your code out for others to use, dont tout a faulty product if you have a fixed version.

I will not be witholding source for months at a time, that may be my choice but I'd like to keep everything as open and friendly as possible. Competition taking code shouldnt be reduced by keeping source udner wraps, ti should be avoided by making your AI more distinctive, adn building the AI from a better Framework.

An example of this would be the NTAI way fo building versus the JCAI/AAI/OTAI way fo building. NTAI's way doesnt dither like JCAI's does, but JCAI is mroe flexible and robust, NTAI's a rigid build tree. Both ahe their advantages and not becuse of little btis of code btu structure and uniqueness.

I too have started on finishing the transformation fo the AAI code into the Map and grid sector system I outlined in NTAI X, both systems fit nicely into the same category and arent too hard to switch between.

Submarine, your AI is where it is because it is easily udnerstandable and has an elegance of simplicity about it. It is that you gave some of us things we needed and filled a niche. AAI needs to be aggressive, but its itnernals are a godo example of how to do things, and have a good future in an aggressive expansionist AI.

Spring RTS Engine

The Opensource paradox