Some semi-random thoughts on terrain generation
We need the plugin first. And I'm not sure if RasPi is a good platform for this, we might need a bit more memory to support potentially more players.
Reply
Thanks given by:
I can run it forever on a 128mb ram and 256mb vSwap vpsTongue.
If the world is limited and view distance is 4 it could handle 10 players or more if you fix player join/leave crashes.
Reply
Thanks given by:
About the crashes, I haven't seen it crash after the ByteBuffer off by one fix. The spamming of join or leave messages is still there though.

Similar to toni, there is a MCS in my signature too.
Reply
Thanks given by:
I'm currently running MCS on an Amazon AWS instance. It's the only one you can get for free (if you just have one of them), and I think its pretty small when it comes to hardware specs.
Reply
Thanks given by:
This has gone quite off-topic. Let's stop listing our servers, we'll discuss that when the time comes.

I think I have an idea about how to generate the heightmap for the mesa bryce biome. It's quite simple, actually Smile
We'll need two Perlin noises, one with low frequency and small amplitude, the other with medium frequency, high amplitude and very low basepoint. Then for each column just take the bigger of the two values. The low-frequency noise will produce the red-sand floor, and the high-amplitude one will punch out the towers from time to time.
[Image: mesa-bryce.jpg]
Reply
Thanks given by:
But if you use two noises wouldn't it slow down the generation speed?
Reply
Thanks given by:
I don't think it'll matter that much. But we won't know until we try Smile
Reply
Thanks given by:
Sounds like it would work fairly well.
Reply
Thanks given by:
xoft, regarding this:
[Image: mesa4.jpg]
I think it might benefit from some low-freq low-amp waveform for "shifting" lines up and down.
Reply
Thanks given by:
Also xoft, eset makes branded square paper???
Reply
Thanks given by:




Users browsing this thread: 3 Guest(s)