Just a Gwai Lo fun within prescribed limits [?]

Back in August I mentioned that I had started reading Lifebox, the Seashell, and the Soul: What Gnarly Computation Taught Me About Ultimate Reality, the Meaning of Life, and How to Be Happy by Rudy Rucker and have, since then, been quietly posting chapter notes from the book. A little more dense than I expected, which should become clear when reading my notes, but enjoyable still.

This is not the first time I posted chapter notes of a book while still reading it: back in 2004, for Urban Vancouver, I posted chapter notes of The Corporation by Joel Bakan. In this case, keeping with the permanently experimental nature of Just a Gwai Lo, thanks to a newer version of Drupal combined with the Views module, my chapter notes almost looks like a separate weblog about the book. Though conceivably I could have done it with a tag (like 'lifebox') an outline with a view made a little more sense, since I can use the tag for links to external articles about the book or photos of the book. Someone already has, in fact: at this writing, the only two photos tagged with 'lifebox' at Flickr are photos using the book to display bookmarking technique: 1, 2.

Rudy Rucker's third chapter of The Lifebox, The Seashell, And the Soul tackles evolution as computation and the possibility or artificial life (which he shortens to a-life). Of the three so far, this is the most accessible, talking about activators and inhibitors (relating them to ideas and their spread), and later in the chapter, cellular automata algorithms used to produce what look like trees. Most compelling, however, was his discussion on search method algorithms, though as a computer geek, I had to remind myself that he did not limit himself to search engines as we know them on the Internet but the ways and the feasibility of the ways we as humans and potential artificial life system would search for the solution to a problem. He distinguished between best solutions and acceptably good solutions, because it's far more efficient, not to say worthwhile, to find the acceptably good solutions over best solutions, because the later most often means you have to search through all possible solutions:

The fact is, given a large enough search problem, none of our various search methods is likely to find the absolute best solution with a feasible amount of time. But so what? After all, absolute optimality isn't really so critical. In any realistic situation we're quite happy to find a solution that works reasonably well. Nature is filled with things that work in clever ways, but it's a delusion to imagine that every aspect of the biome is absolutely optimal. Surely we humans don't have the absolute best bodies imaginable, but—hey—our flesh and bone holds up well enough for eighty or so years. We do better than our rival apes and our predators, and that's enough.

Nearer to the end he quotes at length from his novel The Hacker and the Ants, about a scenario in which software developers create virtual worlds in which their simulated robots compete amongst themselves in virtual environments. I imagined several homes in The Sims, with robots instead of humans, evolving in houses each slightly different but covering all the possibilities of combinations within those homes, the robots having descended into complete anarchy. I haven't read any of Rucker's novels but they can be found at the local used bookstore, so it's about time I got my hands on a few, since if he uses scenes from them to prove a point, he must think that they have stood the test of time reasonably well.

Rudy Rucker explains three world-views of physics in the second chapter of The Lifebox, The Seashell, And the Soul: the mathematical physics view, the continuous cellular automaton view, and the particle system view. I sorted the three from the view I most understand which is what they taught me in high school to the view I least understood, the view they tried to teach me in university, which forever soured me on the subject. (As for celular automata, they didn't even try to teach us that in school!) The ramp from discussing cellular automata in the book is pretty steep, so I could probably use a book-length treatment of the subject with a nice, gradual incline.

Earlier in the chapter, Rucker defines analog and digital not in terms of absoultes, but in terms of the size of discrete possibilities. His definitions I find satisfying, adding nuance to the definitions of them I had in my head, wihch were something along the line of digital == ones and zeros and analog == fuzzy values. Another satisfying definition—or rather, illustration—comes when he discusses chaos and the effects of small differences from experiment to experiment lead us to conclude that we have to think in terms of averages and probabilities and not absolute laws when predicting events in our physical world. But just as the honours physics class I mistakenly took at SFU lost me with quantum physics, so does Rucker, with photos behaving as a wave and particle based on how one observes the behavior.

The discussion then turns to theories of reality, and Rucker seems to be arguing that time (such a "past" and "future") are an illusion—lunchtime, said Ford Prefect, doubly so. Cause and effect don't so much not exist but rather causes and effects send messages to each other forward and backward through time. This theory, if I'm summarizing it accurately, doesn't get Rucker into the Wikipedia page on the philosophy of space and time, but at least it gives people like me who need an excuse not to attend a meeting in the future by saying we are living in an 'eternal now'. Or something: the argument relies heavily on the scenario presented on pages 138-139 which don't make a lick of sense to this political science major.

In the first chapter of The Lifebox, the Seashell, and the Soul: What Gnarly Computation Taught Me About Ultimate Reality, the Meaning of Life, and How to Be Happy, Rudy Rucker, science fiction writer and computer scientist sets out his argument, which, handily, is the title of the chapter, "Computation Everywhere". The chapter is part organizing—in which he lays out the structure of the rest of the book—and part introductory, defining computations as “a process that obeys finitely describable rules” and classifies them at least two ways, on feasibility and unpredictable lines, as well as how the computations go about. Along the first, computaions can be:

• feasible and predictable
• fessible and unpredictable
• unfeasible and predictable
• unfeasible and unpredictable

and the second, they can be classified “four main behaviors for arbitrary computations that are left running for a period of time”:

1. enter a constant state
2. generate a repetitive or nested pattern
3. produce messy, random-looking crud
4. produce gnarly, interacting, nonrepeating patterns

He then goes on to define terms such as the Principle of Computational Equivalence, the Principle of Computational Unpredictability and universal computations, moving on to analytical machines. The word "machine" in Turing Machine tricked my non-theoretical non-computer science into thinking it was necessarily a physical thing rather than a mathematical or algorithmic concept. Emulation—and that any computer can, theoretically, emulate another—is a concept I have a fairly easy time grasping, if not so much later concepts such as cellular automaton (this time the word cellular tricking my brain into thinking about biological—i.e. existing in physical form—cells), and I'd probably have to see the cellular automaton rules play out in an animation, if maybe sped up.

In subsequent chapters Rucker applies computations to biology and physics, the latter I did well enough in high school to be able to follow along and the former I have about zero knowledge of. The book is so far well-written enough that I can plod through and jump back or pause to research more detailed explanations, though, so plod on I will.