README v4.0

README.md

@@ -1,48 +1,64 @@
 # Arkonia
-A natural selection simulator
 
-[![GitHub release](https://img.shields.io/github/release-pre/SaganRitual/Arkonia.svg?style=plastic)](https://github.com/SaganRitual/Arkonia/tree/v3.0)
+[![GitHub release](https://img.shields.io/github/release-pre/SaganRitual/Arkonia.svg?style=plastic)](https://github.com/SaganRitual/Arkonia/tree/v4.0)
 [![License](https://img.shields.io/github/license/SaganRitual/Arkonia.svg?style=plastic)](https://github.com/SaganRitual/Arkonia/blob/dev/LICENSE)
 [![Runs on](https://img.shields.io/badge/Platform-macOS%20only-blue.svg?style=plastic)](https://www.apple.com/macos/)
 [![Does not run on](https://img.shields.io/badge/Platform-not%20iOS-red.svg?style=plastic)](https://www.urbandictionary.com/define.php?term=SOL)
 
-![Awesome gif of Arkonia missing](https://github.com/SaganRitual/Arkonia/blob/exp/MovieforREADME.gif)
+## A natural selection simulator
 
 A hobbyist's toy, under construction. A world of creatures whose brains and bodies evolve
 genetically to make them better at eating each other, and ultimately, to make you all my slaves.
 Written in Swift, using SpriteKit, with no UI other than the main window, because I really
 suck at UI programming.
 
-# Status 09 April 2020
+# Status 10 April 2020
 
-## How to Build n Stuff
+## Now We're Somewhere
 
+<img align="right" src="https://github.com/SaganRitual/Arkonia/blob/exp/MovieforREADME.gif">
+
+At long last, it's fast enough and stable enough to be (potentially) interesting.
 I notice that a few people have cloned the repo. That's cool. I do most development on the `exp`
 branch. I merge down to `dev` whenever I get sick of chasing bugs long enough to do the merge
 and push. So look for the latest on `exp`. Build the `Karamba` Xcode project, debug or release
 
 ## The Latest on The `exp` Branch
+
 ### Metal and Accelerate support for neural nets
 
-Now we can use the BNNS functions from the Accelerate library, in addition to the Metal libs for
-using the GPU, as well as the BLAS stuff in Accelerate. As it turns out, Accelerate does just as well
-as the GPU; I think it's because the nets are too small to be efficient. Some investigation is in order,
-but I haven't done it yet. Also, I think I broke the GPU and BLAS stuff while getting the BNNS
-stuff to work
+I've created neural net drivers for both the `BLAS` and `BNNS` models in the Accelerate library,
+as well as drivers for the GPU using both the matrix operations and the convolution kernel tools
+in the the Metal Performance Shaders library. As it turns out, Accelerate does just as well
+as the GPU; I think it's because my neural nets are (currently) too small to be efficient.
+Some investigation is in order, but I haven't done it yet. Also, I think I broke the GPU and BLAS
+stuff while getting the BNNS stuff to work, but now that I know so much more than I did then, I'll
+probably rewrite those rather than trying to fix them
 
-### Pollenators
+### Selection pressures
 
-Mysterious disc-shaped creatures that float around in the Arkonian ooze, pollenating the manna
-so it will regrow after it has been grazed
+* The obvious one, finding the manna (food): Arkons don't know how to do that when the app launches
+* Pollenation: the manna can't re-bloom after being grazed until it's pollenated. The pollenators
+are mysterious, expertly-drawn, disc-shaped creatures that float around randomly in the Arkonian ooze
+* Manna fatigue: manna takes longer to re-bloom to full nutritional value when it's overgrazed; it
+needs time to recover
+* Oxygen: there's a trace amount of oxygen in the muck, which arkons can absorb only when they move;
+most of what they need is in the manna, so they kind of have to eat constantly
+* Per-neuron energy cost
+* Per-mass energy cost
+* Gestation and birth costs
+* Entropy
+* Carbon-dioxide reducer or whatever-something: some random experiments with other selection pressure,
+I'll clean it all up some day
 
 ### Sensory inputs
 
 Arkons currently have four sets of senses, two sets for tracking external stimuli and two
 for tracking internal stimuli:
 
+* Location in the world, as (x, y)
 * Hunger level
 * Oxygen level
-* Location in the world, as (x, y)
 * Variable-range detection of food and other arkons (which can be eaten if the attacker is big
 enough), as ((x, y), nutrition value) pairs
 * Locations of pollenators, as (x, y)