My problem (well, one of them, anyway) is that I like to experiment. This is great and all, but it means that I end up with a lot of half-completed experiments and one-off demos sitting around, that don't do me a whole lot of good. A lot of things never reach "official" status. For example, I've done a lot of tinkering with procedural textures and so forth, and I've done a lot of demos where I'll render out the results of an experiment and go "ooh, pretty", then lose/erase/change things and never see that again. Then when it comes time to do an actual texture for an application, I'll think to myself "That one texture would be perfect; too bad I have no idea how I made it." Then I end up trying to cobble together something from scratch.
Well, recently I've done some work along the lines of procedurally generating the procedures to generate textures; I write a recursive function that can generate a tree of noise modules from which to generate a texture. I've gotten some pretty cool textures this way; but, again, it's all been one-off experimental stuff, and dozens of nice looking patterns have been lost. So today I finally implemented functionality to save and store randomly generated function chains in a library system. I generate the table descriptor from which a module tree is built, and store that descriptor in a directory along with a sample rendered image of the associated texture. I implemented a new module type, "library", which rather than instancing a specific module, will instead look up a descriptor in the library and insert the code for the referenced module chain.
So now, I can iterate a script that generates randomized patterns, then stashes the module descriptors in a library. When I start work on a texture, I can just browse the library until I see something I like, then include it into the module descriptor for the texture. I've had a background process running for a couple hours now, and I've gotten some pretty cool looking textures. And better yet, I "remember" them all and can use them again as needed.
Previous Entry
Parse this
Next Entry
Indie Game Graphics on the Cheap
Advertisement
Latest Entries
Mechanics of Flow in Gold Rush.
7022 views
Gold Rush
5514 views
DungeonBot3000: Inventory and Items Display
5812 views
DungeonBot3000: Postmortem
5728 views
DungeonBot3000: Boss Fight Scripting
5129 views
DungeonBot3000: Friggin Laser Beams
5714 views
DungeonBot3000: Video
4850 views
DungeonBot3000: Stats, Combat, and Items
6910 views
DungeonBot3000
5056 views
Diablo:Immortal
13973 views
Advertisement
Advertisement