Thank you ghostwheel!

Thank you for the welcoming

Michaël, please do!

Marco, yes the "behaviours" are stored within the genome.
We store both the shape and the sensors/effectors(actuator) within the genome. Each effector (e.g. muscle on articulation, gripper, etc..) does have one or more functional programming trees that take values from the sensors and do something with. Both the shape and the "brain tree" can evolve by mutations/new generation.

Usually we were evolving either one or the other, because both in the same time create a search field too wide for the genetics algorithms to work efficiently.

To get something more interesting with a quicker evolution schema that'd fit game environment. To let each articulation evolve is not a good solution and you need some building blocks (usually your right and left muscles does have the same kind of mechanism) but to implement some kind of high-level behavior is not a problem and it's possible to make them evolve and adapt to the shape you design.

I had the same deception by playing Spore, and to be able to change drastically the shape of your creature during evolution process was a bit disturbing.
In that case, if the stats were that important, I'd prefer to have control on the stats and let the shape evolve by itself. Discovering interesting effects between characteristics and shape. (e.g: effect on creature's height of having big/multiples eyes => big? or small? creature) A bit like playing with the DNA directly.

Hi notgames forum,

My name is Cyprien, I'm 25 years old and I live in switzerland. I recently attended to a conference featuring Michaël and Auriea from Tale-of-Tales and discovered this community. And I'm really happy/excited to be here !.

Please apologize my approximate English, I'm a french speaker and a bit afraid that my accent can be heard through my words.

About myself: I come from the technical side. I'm a quite efficient computer scientist and I'm currently doing a Master while working a 60% in the industry as a software developer (for http://www.routerank.com). Videogames take a big place in my life, playing them as child on the mac of my father and then still playing them a lot, sometime they haunt my dreams, sometime they help me escape reality and I've always been drawing concepts or ideas on a book while taking the train to get home. I learned to program while doing some kind of webgame (completely unreachable goal was to do a "web based mmorpg"). I made a couple of versions of it, mostly to try-out new technology and technical approach while developing the story and ideas on paper. I never achieve to reunite both of them, but it was a very interesting experience (I was quite young, 15-16 years old) some ruins can still be seen here : http://www.resurrection-jdr.com, but it does not work anymore :p (I'm even surprise that the homepage show up at all).

Then I started a Bachelor in computer science and improved a bit my technical side. For my Bachelor thesis, a colleague (Aurélien Da Campo) and I looked deeper into genetics algorithms and bio-inspired mechanisms. We made a small experimental software, some kind of laboratory that allows creation/edition/simulation and evolution of 3d virtual creatures (can be downloaded and seen here: http://www.genecraft-project.com). The goal of the software is to "design" a creature skeleton composed of bones and articulations and with the help of artifical selection, mutations and criteria of evolution (such as movement speed, jumping height or even number of bones): to let the shape and controller evolve (a lot like Karl Sims work (in 1994) http://www.youtube.com/watch?v=JBgG_VSP7f8).

The piece of software we made is not a game, it's mostly a scientific application with a very little bit of 3d-rendering (however our supervisor's child said that "it is ''better'' than Spore because you can throw block of rocks at creature by right-clicking" ).
But I'm really fan of what we experimented with these bio-inspired techniques. The game industry doesn't use them at lot, because they are hard to setup, somewhat unreliable, less efficient and far more complicated than the well known A-star or old good fashioned AI. However I deeply think that they offer interesting potential regarding data and visuals generation, user interaction, AI and mechanism. I'm currently developing some prototype on that field and I'm looking at any experience in this area you might have.

I never pushed the reflection a lot on why I want to do games and on what kind of games I want to do. But it's a very good question, and I'm glad this place and you made me think about it. To be honest I think that I did some kind of reflection on this subject when I was 16, but I wasn't really ready and complete enough for it, and to be still honest I'm not sure that I am now . But since the conference I'm thinking a lot, writing small notes and more motivated than ever, which sound good . Ideal future personal plan are: finishing my master and starting a small videogame studio. May sound a bit naive, but if I don't at least try, I'll regret it deeply . I'm really looking forward to exchange with all of you, trying to think a bit outside the box and reflect on my motivation behind this life project. In the other hand I offer any technical knowledge that I have if it can help you guys to work on some of the cray ideas you might have.

To end this presentation there's some of the project I'm working on right now:
* Genecraft-project: 3d creature simulation and evolution based on fitness criteria and genetics mutations. http://www.genecraft-project.com
* to be named: Using neural-network to reproduce user reaction to allow his decisions/actions to be continued while doing something else. You might see it as a kind of "RPG" where evolution would be what the application learn from the user and overlapping/combining these reaction together. Best image I have is like having several mouse/input to control something.
* to be named: reflection on reproduction of feeling/mechanism of games such as the one where you have a post-it on the head and with the help of other people you have to guess the content of it (e.g. "Am i a woman?", etc.) but in a more abstract/color/shape/movement/reduced communication way.
Maybe some kind of "Turing test" variant where you have to look like a bot for others players and where your goal is to unmask real players among bots. So you need to get intelligent reaction from them while not looking too smart.
* ...  i'll maybe stop here and read some of the actual discussions and posts of the forum

See you,
Cyprien