I can see that you need LOD for something like a flight-sim, where the user can look down at a large area. For other games I think it's probably nice to have, but not essential. Check out this tree model
for example. It's 324 polygons, apparently. Now imagine you built a scene containing 200 trees, and let's say that the player can see at most half of them at a time. That's 32,000 polygons: peanuts even for a Pi.
Have you tried the Unity engine? It does the same trick with a single plane, and you can get some really strange effects if it's not tuned properly. You walk up to a tree and it kind of stretches itself, as the plane turns into a 3D object. I imagine they try to avoid jumps by switching from the plane to a squashed 3D model, then gradually filling it out. Fine if it's far enough away that you can't see it, but very odd if you're standing underneath it at the time.
PeteX is working away trying to look like he's doing something useful.
I keep myself entertained, anyway!
(I'll post an update at the end of the message.)
Glad to hear the wolf is still going well.
by monsters do you mean harmful animals or are you looking for a few more make believe type things as well
That's a hard one to answer. Remember, we are making a kit for creating multi-player games, not a multi-player game as such. My interest is dark fantasy, so from my selfish point of view I would say yes—zombies and ghosts and wraiths and skeletons please! But that's for the world I want to create using the kit. The kit itself, I suppose, needs a variety of art, to help people get started with whatever kind of world they want to create. Dragons, trolls and giant wasps would all be good for that!
Mayang.com looks pretty good, and their licence seems suitable for open source. That's a good catch—a lot of these sites use licences which don't work for products that can be redistributed by the end users. (If we start making use of a lot of their textures, I might send them an email just to check they are happy with this, as they don't seem to have thought about it when writing their licence.)
Thanks for the pull request for the tree. I think perhaps I didn't give the best advice when I suggested making a lollipop shape. Looking at the TurboSquid tree
, the artist seems to have created a number of planes which have images of branches and leaves UV-mapped onto them. By arranging these planes in 3D space, he created the appearance of a complete tree.
Do you think it would be worth trying to do something similar? It seems to have worked well for the TurboSquid tree, perhaps it can work well for us too. (BTW, if anyone is wondering, we can't just bung TurboSquid $20, and use their tree. Their licence is definitely incompatible with open source.)
Thanks for the interest! What would you be interested in working on? Are you more inclined towards art or code, first of all?
We now have terrain with hills. You can also create a stencil texture, which allows you to choose different surfaces for the terrain. In the terrain I have made, you can choose between bare earth and grass, but of course you can change those textures for other things. You could have snow and rock, or whatever fitted with the landscape you were making.
I've got one more thing to do on this. I need to include the terrain and terrain textures in the list of things that get downloaded from the network when the client is started. Once that is done, I'll push the revisions and those of you with access to Linux will be able to give it a try.
I then need to do the same thing for the skybox, so you get a sky texture rather than a black background.
After that, I want to change gear a bit. I want to produce a version that runs on Windows, for those of you who don't have easy access to a Linux machine. We also need a website with our own forums and things, because the project is growing and we can't continue to coordinate things using a single thread on the Raspberry Pi board.
If anyone wants to be involved in the project and doesn't currently have a role, perhaps you could consider doing the website design. That will speed things up; I can set the site up with a simple and minimalist design, then we can upgrade to something nicer later.
Finally, I wanted to ask for people's opinions about the way the toolkit should be customised. How do you turn a general purpose multi-user game toolkit into a particular game? My idea is that you put an entry in the server config file which tells it to load a particular Python module. This happens when the server starts. That Python module then registers interest in various events, such as new players joining the game, players pressing certain keys, players going to particular locations, and so on.
When the customisation code gets called, it can do things like creating objects, moving objects, showing visual effects.
If people are happy with this structure, I will start by creating a few events and API calls. Eventually we'll have to think what the complete list should be, but that's a long way off!