Icon Quaternion Weighted Average


Icon BVHView


Icon Dead Blending Node in Unreal Engine


Icon Propagating Velocities through Animation Systems


Icon Cubic Interpolation of Quaternions


Icon Dead Blending


Icon Perfect Tracking with Springs


Icon Creating Looping Animations from Motion Capture


Icon My Favourite Things


Icon Inertialization Transition Cost


Icon Scalar Velocity


Icon Tags, Ranges and Masks


Icon Fitting Code Driven Displacement


Icon atoi and Trillions of Whales


Icon SuperTrack: Motion Tracking for Physically Simulated Characters using Supervised Learning


Icon Joint Limits


Icon Code vs Data Driven Displacement


Icon Exponential Map, Angle Axis, and Angular Velocity


Icon Encoding Events for Neural Networks


Icon Visualizing Rotation Spaces


Icon Spring-It-On: The Game Developer's Spring-Roll-Call


Icon Interviewing Advice from the Other Side of the Table


Icon Saguaro


Icon Learned Motion Matching


Icon Why Can't I Reproduce Their Results?


Icon Latinendian vs Arabendian


Icon Machine Learning, Kolmogorov Complexity, and Squishy Bunnies


Icon Subspace Neural Physics: Fast Data-Driven Interactive Simulation


Icon Software for Rent


Icon Naraleian Caterpillars


Icon The Scientific Method is a Virus


Icon Local Minima, Saddle Points, and Plateaus


Icon Robust Solving of Optical Motion Capture Data by Denoising


Icon Simple Concurrency in Python


Icon The Software Thief


Icon ASCII : A Love Letter


Icon My Neural Network isn't working! What should I do?


Icon Phase-Functioned Neural Networks for Character Control


Icon 17 Line Markov Chain


Icon 14 Character Random Number Generator


Icon Simple Two Joint IK


Icon Generating Icons with Pixel Sorting


Icon Neural Network Ambient Occlusion


Icon Three Short Stories about the East Coast Main Line


Icon The New Alphabet


Icon "The Color Munifni Exists"


Icon A Deep Learning Framework For Character Motion Synthesis and Editing


Icon The Halting Problem and The Moral Arbitrator


Icon The Witness


Icon Four Seasons Crisp Omelette


Icon At the Bottom of the Elevator


Icon Tracing Functions in Python


Icon Still Things and Moving Things


Icon water.cpp


Icon Making Poetry in Piet


Icon Learning Motion Manifolds with Convolutional Autoencoders


Icon Learning an Inverse Rig Mapping for Character Animation


Icon Infinity Doesn't Exist


Icon Polyconf


Icon Raleigh


Icon The Skagerrak


Icon Printing a Stack Trace with MinGW


Icon The Border Pines


Icon You could have invented Parser Combinators


Icon Ready for the Fight


Icon Earthbound


Icon Turing Drawings


Icon Lost Child Announcement


Icon Shelter


Icon Data Science, how hard can it be?


Icon Denki Furo


Icon In Defence of the Unitype


Icon Maya Velocity Node


Icon Sandy Denny


Icon What type of Machine is the C Preprocessor?


Icon Which AI is more human?


Icon Gone Home


Icon Thoughts on Japan


Icon Can Computers Think?


Icon Counting Sheep & Infinity


Icon How Nature Builds Computers


Icon Painkillers


Icon Correct Box Sphere Intersection


Icon Avoiding Shader Conditionals


Icon Writing Portable OpenGL


Icon The Only Cable Car in Ireland


Icon Is the C Preprocessor Turing Complete?


Icon The aesthetics of code


Icon Issues with SDL on iOS and Android


Icon How I learned to stop worrying and love statistics


Icon PyMark


Icon AutoC Tools


Icon Scripting xNormal with Python


Icon Six Myths About Ray Tracing


Icon The Web Giants Will Fall


Icon PyAutoC


Icon The Pirate Song


Icon Dear Esther


Icon Unsharp Anti Aliasing


Icon The First Boy


Icon Parallel programming isn't hard, optimisation is.


Icon Skyrim


Icon Recognizing a language is solving a problem


Icon Could an animal learn to program?




Icon Pure Depth SSAO


Icon Synchronized in Python


Icon 3d Printing


Icon Real Time Graphics is Virtual Reality


Icon Painting Style Renderer


Icon A very hard problem


Icon Indie Development vs Modding


Icon Corange


Icon 3ds Max PLY Exporter


Icon A Case for the Technical Artist


Icon Enums


Icon Scorpions have won evolution


Icon Dirt and Ashes


Icon Lazy Python


Icon Subdivision Modelling


Icon The Owl


Icon Mouse Traps


Icon Updated Art Reel


Icon Tech Reel


Icon Graphics Aren't the Enemy


Icon On Being A Games Artist


Icon The Bluebird


Icon Everything2


Icon Duck Engine


Icon Boarding Preview


Icon Sailing Preview


Icon Exodus Village Flyover


Icon Art Reel




Icon One Cat Just Leads To Another

How Nature Builds Computers

Created on April 23, 2013, 5:32 p.m.

At a basic level life could be seen as any self-replicating system. A physical system in the universe which manages to remain stable for long enough that it can use the resources around it to create replicates. We stack many other qualities for life too, but self-replication is the ultimate requirement. Otherwise it is nothing more than a chemical happenstance in the vast universe.

Computer Shell

At first this self-replicating biology may have happened by chance, like some form of chemical chain reaction or equilibrium. Systems with a miniscule lifetime, and no real ability to fight against entropy and change. Complicated behavior is required if these systems are to last more than a few seconds. Movement and feeding, some protection against the elements of the outside world.

Fighting against the weathering and unpredictability of the environment is the biggest challenge to life. On our blue planet evolution has stepped up to this challenge. It has created life that is incredible and diverse. There are many lessons to learn from it. Not just lessons in creating stable systems, but creating beautiful ones too. The scorpion is one of my favorites.

When we create computer programs meant to be left to their own devices, we are in a sense creating these autonomous systems, much like animals operating in a simple environment. Can we analyze Mother Nature to see how she manages to adapt so well to so many complex and unpredictable environments?

One approach is to take conservative action, observe the environment carefully, and only act when there are predictable outcomes. Another is to use formal verification or powerful simulation to prove that no components of the system can fail in any situation. But I am talking about something more fundamental and devoid of these high level mathematical and logical concepts.

A radical idea comes from scientists, which in 2010 compared the functionality of an E. Coli bacteria to that of the Linux Kernel. They compared the "call graph" of the systems.


Call Graph


In an E. Coli there are a few basic functions that can be performed at the very top level. These are the overall states, or behaviors, of the bacteria. These all use some middle level systems and controllers, which then use a very large number of different basic physical functions for creating proteins and enacting biological needs.

The E. Coli has a few simple goals in life. A few simple high level things it tries to achieve. But in trying to achieve these there are a large number of logical and physical processes going on at the low level. This muddle usually does what is required by the high level goals - but the method is convoluted and somewhat imprecise. It is no design any sensible person would use.

The Linux kernel is the opposite. At the top level is a number of user programs that perform all the tasks needed by a modern computer. They call on many different mid-level controllers which manage the major systems such as networking, audio and video. These finally call a small set of important functions which run the core of the machine. The Linux kernel is imprecise and fuzzy at the high level - while well defined and simple the closer it gets to the machinery.

In this way the systems are somewhat reflected vertically. The E. Coli resembling a pyramid and the Linux Kernel representing an upside down pyramid.

Their reliability is also mirrored. In the Linux kernel it is most important that the core functionality at the bottom always performs correctly and can be relied on. Otherwise all the other systems break. At the middle level it is still important that everything functions, but one thing breaking does not make the system completely dysfunctional. At the top programs often crash. While annoying there are usually other ways to achieve the same task.

In the bacteria evolution has rendered the bottom layer somewhat unpredictable. Some of the basic functions don't work, some only work occasionally and some do something only similar to what was intended. The middle layer which manages behavior is also unpredictable, but will tend toward the correct behavior. The top layer is most important for a correctly functioning organism.


Computer Knobs

The reason this backward design is so successful in nature is due to mutation. Mutation always happens at the bottom most level - the genetic machinery - the DNA. If a fundamental component were to change at the bottom level in the Linux kernel the whole system would be broken. This isn't true in the bacteria. A break in one bottom level piece of machinery doesn't make much difference. There are so many other routes to perform behavior that it will probably still function the same.

These design schemes are reflected in both all of nature and computer science.

In computer science we love perfect abstractions. Those components we can use without knowing the internals, and that we can use without them ever unexpectedly failing. The computer in my mobile phone has been clicking over for five years and never got an addition wrong. Another example is the internet. One can use it without any idea of how it works. It is vast and complicated, but (almost) always just works.

In nature things are softer. Animals tend to have a few simple behaviors. But the systems which carry those behaviors out are huge and complex with many flaws and faults. They have simple, unreliable ways of overcoming those issues. Somehow it almost always works out in the end.

To the guided hand of a watchmaker the upside down design of an E. Coli is abhorrent. The complexity is intractable and chaotic. It is the design of a Jackson Pollock madman.

But Mother Nature is no mere watch maker. Evolution laughs at a system that has run for just five years. There is another ledge to climb onto in system design. Eventually programmers must assign her role. The computers will have to start doing the designing for us. And they will do a horrible job of it. But somehow it will feel more familiar. They will have the quirks of the flora and fauna we love. Somehow those systems will start to feel alive.



github twitter rss