confocaloid wrote: ↑February 5th, 2024, 11:30 am
I think I would like to know what is the intended target audience for those pages. They contain lots of formulas, with little or no explanation as to why all this belongs to LifeWiki. Some parts look like SPAM. Are you writing for other people?
As an example, the edit
https://conwaylife.com/w/index.php?diff ... did=144364 added a new section with content collapsed by default. (Collapsing content by default is already a bad enough idea, from the accessibility viewpoint. If you don't want to keep the content visible by default, then it is probably better to avoid putting that content on that page.) The hidden content is basically spam by formulas without explanations why these are in the LifeWiki scope.
I think most of those additions should be removed, as out-of-scope (and hard to read) content added by an overenthusiastic editor.
I don't believe changing formatting of expressions could automatically make the pages easy to read. There should be much more explanatory text written for readers.
And of course changing the formatting cannot solve the problem with content that simply does not belong on LifeWiki.
I have mainly been adding the results I find without much in the way of explanation, I confess, however I intend to revise it into a more useful explanatory page eventually (after finishing my current addiion, I will write up a proof of the main claim (the bitwise recurrence relations) that will be more plain English than the existing proof of its solution). The new part containing generating functions is there because the sequences l(t) and v(t) are given by quite complicated forms, but become simpler when expressed through their g.f.'s (since forms involving infinite products of rational functions correspond with recursive forms for their series expansions), and I wanted to include those of all of the others for completeness, for all that it was worth.
Also, finding generating functions for sequences of on-cell-counts in cellular automata is not actually a very esoteric idea! The Sierpinski triangle is well-studied and the simplest case, another is rule 150, and a few 2D rules have been studied also
- N. J. A. Sloane, On the Number of ON Cells in Cellular Automata, 2015
- Shalosh B. Ekhad, N. J. A. Sloane, Doron Zeilberger, A Meta-Algorithm for Creating Fast Algorithms for Counting ON Cells in Odd-Rule Cellular Automata, 2015 (note that odd rules here do not refer to their rule integers being odd, but to those which (alike 150) are distributive with XOR, since they can be considered to enumerate odd terms of expansions of powers of polynomials)
- Alan J. Macfarlane, On generating functions of some sequences of integers defined in the evolution of the cellular automaton Rule 150, 2016 (this one doesn't seem especially well-written to me, and is somewhat convoluted)
It seems rule 120/225 was overlooked by many people for the 42 years since its discovery (I suspect because Stephen Wolfram had already found the main result result regarding the domino's growth rate, and one would either have to consider the odd member of its equivalence class or the even member from two cells to even see the nontrivial behaviour), even though it is the only range-1 rule that produces this fractal.
I asked dvgrn about it over forum mail, and he said it was okay and he in fact welcomes the creation of pages in the OCAspace that are more specific in nature compared to the more introductory pages on the LifeWiki.
No objections from me! The OCA namespace is definitely allowed to be more experimental. It's much more likely to contain documentation of users' own discoveries, mostly just because there are so many OCA rules that each one might only have one person who has done a "deep dive" into exploring it.
and then over Discord,
Heh, I was a math major in college, so I could technically read all of that -- but I have to say, I'm honestly not likely to take the time. The audience for that level of attention to mathematical detail in the community is ... very small, I would think, especially if there might be other non-LaTeX-equation-heavy ways to say the same things.
apgoucher could no doubt read though that and get something out of it quite easily -- but I'm basically just a "recovering math major".
Some parts of the article are of the kind that would belong in the FORMULA section in relevant OEIS sequences, though often they are too general (I think I am the first to consider the extension that is the basis of the generalisations section, as well), and I want to eventually write it all up into a LaTeX paper.
There are other questions that lead to sequences that can be used to characterise 1D rules, ie. "the number of length-n substrings that have finite predecessors in the rule" and "the maximum period of a length-n oscillatory agar with [cylindrical scrolling/fixed end states]," yujh
proved that the union (over all n) of the sets of periods of oscillators in length-n finite tapes with opposite-state ends in rule 150 is the positive integers, as the proof of omniperiodicity of B34kz5e7c8/S23-a4ityz5k. I would like for other people to add their findings so it is not all my particular domain of knowledge, though it seems there are not especially many interested.