Code center > Darwinbots3
How can I best help
d-EVO:
--- Quote from: Numsgil ---I don't know, you'll have to try it and tell me. Right now the only projects that are 2008 are in the DNA and physics directories.
--- End quote ---
I have tried but im not sure if I did it proply
I am using smartSVN and I think I am doing the right thing.
I set up the directory and every thing and I can use it to find the DB3 directory on the site and it shows the C# symbole.
I click it and it does nothing.
probably did it wrong. not used to this subversion thing. The documentation is obviosly not ment to help dumb asses like me .
Numsgil:
No, you're doing the right thing by telling me if you can't figure it out. I make no guarantees that I explain things well
You have to "check out" a local copy of the repository to your hard drive. Did you get that far? If so, then you have to open solution files (*.sln) into the IDE. Well, you don't have to but that's the best way to work. There's a VS2008 .sln file under trunk\modules\DNA\Sunweaver\
ikke:
The physics module project file is not recognised as a valid project by MS c# express 2008.
I was hoping to find the data warehouse to see and learn. Is any of it already implemented in the C++ port?
Q: do you equate modules to threads? I think the program can definitely benefit from multithreading, but I was wondering what your idea is. I have worked with threads in C++ and understand them, but I have never designed and or implemented one, so it would be a first for me
Numsgil:
--- Quote from: ikke ---The physics module project file is not recognised as a valid project by MS c# express 2008.
--- End quote ---
Damn. Half the reason I upgraded to 2008 was to use the built in unit testing feature, but it's not available in the express edition. That really throws a monkey wrench in everything I wanted to do.
--- Quote ---I was hoping to find the data warehouse to see and learn. Is any of it already implemented in the C++ port?
--- End quote ---
Just to be clear, the C++ version was what I worked on about 2.5 years ago. It was a pretty straightforward port of the VB version, and brought along most of the bugs and limitations of the VB version. The C# version really breaks from the old code base and is its own entity. The fact that I was working on physics for DB3 in C++ shouldn't be confused with when I was working on Darwinbots in C++, they were different projects. (My plan was to write a physics module in C++ and write a wrapper around it for the rest of the C# code, so I could recycle the physics code in other projects of mine, but I abandoned that idea).
So the C# version right now is pretty naked. I have a DNA module that's maybe two thirds done, and a few mostly empty projects that I'm slowly filling out, and that's about it.
So in short, the data repository idea doesn't exist in any version of Darwinbots. I decided to pursue it when I switched to C#, and the C# version is far from done.
--- Quote ---Q: do you equate modules to threads? I think the program can definitely benefit from multithreading, but I was wondering what your idea is. I have worked with threads in C++ and understand them, but I have never designed and or implemented one, so it would be a first for me
--- End quote ---
I'll probably employ worker threads inside modules (or whatever level of the code it is that I do "for each bot, do X"). It's easy enough to set up.
ikke:
Just went over de data depository article. It has some similarities with what I worked on in my university years. I had a central simulation state with various actors taking turns to modify the state.
Just a dump on my thoughts of implementation in DB3:
It pays to think about the organization of data and the way on how to achieve the link between the modules: relatively speaking either the database is going to be complex and the interfaces simple or the other way around.
For now I see three or four object types in the database:
Shapes (including or maybe even solely consisting out of their edges)
Bots (Corpses as a special type)
shots
Are teleports a separate object type or an instance of a shape?
For each of these types some underlying data needs to be stored in the depository. An overview of what springs to mind and where they would be needed:
Bots:
location / shape (collision detection, rendering, volume, DNA)
Connections (ties) interaction between cells, DNA
Mass (physics, DNA)
Velocity (physics, DNA)
Acceleration (physics, DNA)
Refvars (DNA)
Shapes:
Location / shape (intersection / collision, rendering, DNA))
Physical / chemical attributes (DNA, physics)
Shots:
location (collision, rendering, DNA)
Velocity (physics, DNA)
Type / content (physics, rendering, DNA)
Teleports:
To from link (network)
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version