Saturday, June 8
Cellular Automata Lecture
CINVESTAV
Professor: Harold V. McIntosh.




CELLULAR AUTOMATA (19)


May 28

Someone brought in a copy of ''New Science'' so it is finally possible to see what has been going on. Talk about `kilos de papel;' you can't get paper cheaper than by buying the book. And not one single literature citation anywhere, although the copyright page insists that one must rigorously respect the author's intellectual property rights and all his hard work and cite him as he so fully deserves whenever possible. So the world is asymmetric! One might accuse HIM of outright plagiarism here and there (``Spacefillers'' annoys me especially), but the colophon assures us that permissions were duly sought and obtained.

However, the main point is whether Rule 110 is universal or not (we still don't know, although Conway is criticised for having left the universality of Life incomplete), The missing predicate seems to be the interaction of a burst of four carefully spaced EBars with four equally well spaced C's. This is something we will have to verify, even though it really is one of those metaglider-in-a-macrocell situations, and thus may strain the capabilities of our programs. At least mine, although the one the challenge was run on ought to handle it.

We have to look at this in some detail. It is one thing to get a structure A which will block structure X but not Y, and another to install A at the other end of the tag chain. Also there is some discussion of a startup transient so we ought to look over Cook's challenge with all these details in mind. And the question still remains, whether you can set up the initial data without the intervention of a universal computation or not. But anyway, we finally have something concrete to look over and study.

The list of gliders is the one we know, so if he has found any others he hasn't reported them. What's extensible and what isn't seems somewhat muddled to me; it bears checking. And the table of productions still remains incomplete; in fact, he doesn't go into collisions at all except for the ones involved in the universality construction.

As far as tiling, he comes close in a discussion of Rule 60, and he reports some work of Cook on polyominoes, but it doesn't look like he thought of applying it to Rule 110.

The question of reversible rules gets a very short treatment; he alludes to blocks but I didn't see that he equates reversible rules with block permutations in any but the most offhand way. However, the whole text is given to casual assertions that this or that is true, for which the lack of citations offers no solace.

- hvm

Return.