When developing Bots what do you used, ie whats in the code? - Page 2

Pierre-Marie Baty

bleh, have you forgotten the RACC preview ? its navigation was better than those of MANY waypointed bots ! to speak like that you mustn't even have tried it I bet

@$3.1415rin

i still see that bsp approach as some way waypointed ... well, again, the waypoints are named surfaces and they have some sort of relation to each other, otherwise you couldnt run A* on it. so to me that falls into the category 'waypointed' and that's maybe the best way to go.

but we had this already in amsterdam

stefanhendriks

@ pmb, i was speaking in general

Not about bots specific.

I did not try your preview long, i had it on my disk sometime, but as you know i do not test other bots very long at all.

BSP is some sort of waypointed ofcourse, since it uses eventually a 'chain' of information to get to positions (waypoints) to navigate through a map. (As asp said).

Totally unwaypointed is doable i guess, though it also requires some sort of memory system so a bot does not make mistakes often in U shaped rooms...

Lethal

Right now in Algebra 3 (don't know why the course is called that when its has nothing to do with algebra) we are learning about tree graphs, Prim's algorithm.

Lethal

And while take a flick through my printed lecture notes I see something called Eulerian graphs, disjoint cycles, and vertex colouring. I guess I will have to wait for my lecturer to explain this, cause he gives us proofs and background then explains and gives examples during lectures.

Pierre-Marie Baty

@stefan and asp, I was talking of my bot before Amsterdam ; back then I was not using anything waypoint related (waypoints, navmesh, BSP stuff). The navigation was exclusively based on tracelines (FOV scan), and the bot could navigate de_dust and de_dust2 almost as well as the best bots at the moment. But yeah, right now the BSP stuff I'm making is totally waypoint related. It's been a big switch since Tub showed us that.

@$3.1415rin

yes, i know, but back then you were thinking about memory nodes and you refused to see them as some sort of wp system

anyway, back to topic:

@lethal : maybe you could posts links to your lecture notes ( if there are any ) or scripts. might be interesting to take a look at them

Lethal

I hope I don't in bother for this, oh well. He has them set to global so I suppose he doesn't mind anyone outside of uni taking a look at them;
Thats the base web site for students looking for thins to read : http://www.ma.hw.ac.uk/~markl/teaching/F11ME3/
And here is the lecture notes in pdf :
http://www.ma.hw.ac.uk/~markl/teachi...lgebra3lec.pdf

Rifleman

Quote:

Originally Posted by Pierre-Marie Baty

bleh, have you forgotten the RACC preview ?

We didn't forget , but we cant get it :-/ . Maybe upload a copy into the filebase ?

stefanhendriks

Quote:

Originally Posted by Pierre-Marie Baty

@stefan and asp, I was talking of my bot before Amsterdam ; back then I was not using anything waypoint related (waypoints, navmesh, BSP stuff). The navigation was exclusively based on tracelines (FOV scan), and the bot could navigate de_dust and de_dust2 almost as well as the best bots at the moment. But yeah, right now the BSP stuff I'm making is totally waypoint related. It's been a big switch since Tub showed us that.

We should just for fun retry that one time... I think its not impossible to do it via that way. Using some sort of memory system.... so it will probably become some sort of waypointed, yet unwaypointed, navigation system. (since you do not rely on waypoints to create paths entirely... if you get what i mean)