4 ms·

Watching this video makes me wonder if Wolfram's "New Kind of Science" is more worthy of study. There was so much controversy about the book and Wolfram's clai

Watching this video makes me wonder if Wolfram's "New Kind of Science" is more worthy of study. There was so much controversy about the book and Wolfram's claims, that I didn't bother with it.

mck- 14y ago
Watching this video makes me wonder if Wolfram's "New Kind of Science" is getting old (like Tetris). I was actually reading that book at the book store for hours, getting quite intrigued, but at the same time deterred by its size. Perhaps I should give it another shot some day.
- andolanra 14y ago
  The general criticism of Wolfram's book is that it's heavy on pictures and self-praise and comparatively light on content. It's certainly very pretty, and has some interesting results (e.g. that Rule 110 is Turing-complete) but a lot of the work is summary or rehash of other extant (and often non-cited) material. There are often more thorough, more concise, and more lucid resources on cellular automata from other research; I'd cite Zuse's Rechnender Raum (Calculating Space) off the top of my head for a work that comes to many of Wolfram's conclusions several decades prior to Wolfram. In fact, a good strategy for becoming familiar with cellular automata as a field is to read reviews of ANKOS by people with degrees[1], collect the works and authors mentioned, and trawl them. You'll learn a lot about a lot that way. [1]: http://shell.cas.usf.edu/~wclark/ANKOS_reviews.html http://shell.cas.usf.edu/~wclark/ANKOS_reviews.html has an extensive selection of such reviews. EDIT: It's also worthy of note here that Wolfram would reject the original post's model as not worthwhile, as ANKOS explicitly argues that we should reject continuous models in favor of discrete models; indeed, that is the very definition of his "new kind of science."
  - lmm 14y ago
    Note that the proof that Rule 110 is Turing-complete was actually invalid. (Also, like much of the work in the book, it was written by an uncredited employee rather than Wolfram)
    - andolanra 14y ago
      I knew that the proof was actually Matthew Cook's work, but I didn't realize it was invalid. Can you point me to a citation for that?
      - lmm 14y ago
        I was thinking of http://science.slashdot.org/story/07/10/29/2212226/wolframs-23-turing-machine-not-universal http://science.slashdot.org/story/07/10/29/2212226/wolframs-... ; looking again it seems that's a different proof, apologies.
kriro 14y ago
I own the book and have read most of it (not deeply). I'll need to actually study it and have reserved a decent time slot for that next year. I think the concepts are fascinating but the representation is somewhat meh. Wolfram seems to be quite fond of himself. That ebing said I don't understand why people would not want to read the book because they "don't like" Wolfram. I recommend it to anyone really it got me thinking about some pretty interesting ideas. Just be open minded and treat it as a "creative tickler" and not a ridid new science :)

Watching this video makes me wonder if Wolfram's "New Kind of Science" is more worthy of study. There was so much controversy about the book and Wolfram's claims, that I didn't bother with it.

mck- 14y ago

Watching this video makes me wonder if Wolfram's "New Kind of Science" is getting old (like Tetris). I was actually reading that book at the book store for hours, getting quite intrigued, but at the same time deterred by its size. Perhaps I should give it another shot some day.

andolanra 14y ago
The general criticism of Wolfram's book is that it's heavy on pictures and self-praise and comparatively light on content. It's certainly very pretty, and has some interesting results (e.g. that Rule 110 is Turing-complete) but a lot of the work is summary or rehash of other extant (and often non-cited) material. There are often more thorough, more concise, and more lucid resources on cellular automata from other research; I'd cite Zuse's Rechnender Raum (Calculating Space) off the top of my head for a work that comes to many of Wolfram's conclusions several decades prior to Wolfram. In fact, a good strategy for becoming familiar with cellular automata as a field is to read reviews of ANKOS by people with degrees[1], collect the works and authors mentioned, and trawl them. You'll learn a lot about a lot that way. [1]: http://shell.cas.usf.edu/~wclark/ANKOS_reviews.html http://shell.cas.usf.edu/~wclark/ANKOS_reviews.html has an extensive selection of such reviews. EDIT: It's also worthy of note here that Wolfram would reject the original post's model as not worthwhile, as ANKOS explicitly argues that we should reject continuous models in favor of discrete models; indeed, that is the very definition of his "new kind of science."
- lmm 14y ago
  Note that the proof that Rule 110 is Turing-complete was actually invalid. (Also, like much of the work in the book, it was written by an uncredited employee rather than Wolfram)
  - andolanra 14y ago
    I knew that the proof was actually Matthew Cook's work, but I didn't realize it was invalid. Can you point me to a citation for that?
    - lmm 14y ago
      I was thinking of http://science.slashdot.org/story/07/10/29/2212226/wolframs-23-turing-machine-not-universal http://science.slashdot.org/story/07/10/29/2212226/wolframs-... ; looking again it seems that's a different proof, apologies.

kriro 14y ago

I own the book and have read most of it (not deeply). I'll need to actually study it and have reserved a decent time slot for that next year. I think the concepts are fascinating but the representation is somewhat meh. Wolfram seems to be quite fond of himself. That ebing said I don't understand why people would not want to read the book because they "don't like" Wolfram. I recommend it to anyone really it got me thinking about some pretty interesting ideas. Just be open minded and treat it as a "creative tickler" and not a ridid new science :)