Some semi-random thoughts on terrain generation

[FakeTruth]

Haven't tried it yet, I'm bogged down in shift-clicking code. The UI architecture is terrible and the worst part is that I cannot come up with anything better.

[xoft]

was my patch usefull?

Well, not really. You took an existing generator, copied its code and made it into another generator. That's not too useful, because if a bug is found in one, it has to be fixed in both. It'd be much better to modify the generator so that it can create both structures based on its parameters.

Also, I've already written that once, I think dirt and gravel "pockets" need to cross chunk boundaries, so the generator should be a bit different than OreNests - it needs to keep track of "pockets" in adjacent chunks.

I'll still try out world generated by your patch and if it feels okay, I'll work on it somehow to get it in.

[xoft]

It looks different from the vanilla world, but it's not bad. Still, I think it deserves a better implementation, so I've added it to the list of easy tasks included into the OreNest refactoring.
R-T-B said he wanted to join the project and he wants to do terrain generation, so I think he may be interested in implementing this.

[NiLSPACE]

cool the terrain underground is now a bit boring if you explore caves, becouse its only stone (and sometimes some ores)

[R-T-B]

It looks different from the vanilla world, but it's not bad. Still, I think it deserves a better implementation, so I've added it to the list of easy tasks included into the OreNest refactoring.
R-T-B said he wanted to join the project and he wants to do terrain generation, so I think he may be interested in implementing this.

I did a bit of it in Java code, but to be dead honest, I'm not sure my C++ skills are up to par to start on such a large task (and porting say, PhoenixTerrainMods code to C++ would certainly be non-trivial!). It'll be some time before I can tackle that, as from what I've gathered thus far, C++ is a much more "machine level" language than Java. It reminds me of my days of programing in assembly in a sort of oddball way... It's obviously higher level and thus easier to read than that, but still makes my brain tired after a large session. I think I'll get better at it after I wrap my mind around it though.

One benefit I can immediately see is the additional control it gives you over something like Java, which handles pretty much everything low level for you. This is a double edged sword, I'm pretty sure you could code a terrible program or a very fast one, all depending on how lazy you are in your code... Fortunately even in Java I was always using proper variables and such almost to the point of obsessive compulsiveness with memory efficiency, so I should be right at home.

The next thing I need to do after I do some C++ tutorials is actually code on your project in some form... and that probably means I should load up your server and look at the state of things...

[NiLSPACE]

i made a new finisher its cactus for in desert and DesertHills only. there is one problem the SprinkleFoliage finisher still generates Cactus on all sand (also beaches and oceans if there is a piece of sand free thanks to a cave) i was hoping you could delete that in SprinkleFoliage becouse i tried so much but it didn't work for me.

[xoft]

SprinkleFoliage sucks big time, no questino about that, but I'm not gonna change it. There's an expectation that a completely new finisher will be made that will supersede it, and SprinkleFoliage will be left only for worlds where their admins don't want the terrain to change.

I'll see later if the patch is still worth including, and if it is, I'll commit it. But don't get your hopes up

[NiLSPACE]

it's basically a editted version of the lilypad finisher

[NiLSPACE]

was my finisher good enough?

[xoft]

I'm still working on repacketing, so that we can upgrade to 1.3.2 as soon as possible, so I haven't even seen it yet. Sorry.

Maybe someone else could give it a whirl? FakeTruth? Lapayo? RTB?