3 ms·

This has always raised a question to me: >What does an algorithm look like? Is there a real-world possibility of creating something like to _The Glass Bead Ga

This has always raised a question to me: >What does an algorithm look like? Is there a real-world possibility of creating something like to _The Glass Bead Game_ from Herman Hesse's novel (originally published as _Magister Ludi_)? As a visually oriented person, I'd like to think so, and have actually tried to use such tools: https://community.carbide3d.com/uploads/default/original/3X/5/b/5b0872a5666fec9b7bb6fd623c431de03263372d.jpeg https://community.carbide3d.com/uploads/default/original/3X/... but the danger has been, for want of an unambiguous answer to the afore-mentioned question, they always run the risk of: https://blueprintsfromhell.tumblr.com/ https://blueprintsfromhell.tumblr.com/ https://scriptsofanotherdimension.tumblr.com/ https://scriptsofanotherdimension.tumblr.com/ and it's hard striking a balance between visual expressiveness and modularity (which all-too readily results in the wall of text which one is presumably trying to avoid).

boothby 3y ago
> What does an algorithm look like? Writing Piet by hand is a fun way to explore this. Sergei Lewis[1] and I[2] have each written tools to generate Piet code. Sergei's assembler makes much nicer looking code than my Piet backend. All you can really see from my compiler's output is that I'm really lazy and used a trampoline[3] [1] http://www.toothycat.net/wiki/wiki.pl?MoonShadow/Piet http://www.toothycat.net/wiki/wiki.pl?MoonShadow/Piet [2] https://github.com/boothby/repiet/ https://github.com/boothby/repiet/ [3] https://en.wikipedia.org/wiki/Trampoline_(computing) https://en.wikipedia.org/wiki/Trampoline_(computing)
andoando 3y ago
I have a belief that there is a 1 to 1 relationship between any algorithm, or even any mental conception, and a visual representation. I got this idea reading one of Steven's pink book, in which he says that any abstract word can be broken down into simpler words which in the end describe some spatial relationship. For example to "rekindle" is to "bring two things back together". It is easy similarly to say that a for loop is the mental conception of "one thing going over a bunch of other things", which has a visual representation like "100" -> "010" -> "001". I wonder then if its possible to create a language where one defines these constructs purely as visual transformations.
- Pompidou 3y ago
  > any abstract word can be broken down into simpler words which in the end describe some spatial relationship What if my computing model rely on quantum phenomenon ? Maybe some of them don't have good visual representation. But it's a guess. I don't know... Regarding the global idea of your comment: what about languages such as J it APL or BQN ? In such languages each character is an atomic part of algorithms (in j, two characters are used to implement dynamic IF for instance. One character only is used to define the dimensional depth of each part of algorithms). In this way, iversonnian languages can be seen as general and pure visual representation of algorithms once you consider each J or APL character as an arbitrary glyph.
  - andoando 3y ago
    Well consider this: There is some set of words which describe quantum physics perfectly. If what I am saying is true (that human language maps to transformations of space-time) then this theoretical computer should be able to represent quantum phenomena. Im not familiar at all with APL or BQN but I think I understand what youre saying. I think the main point here isnt simply that you can represent programs visually, but that the visuals do exactly what you see. A "combine" instruction for example would literally be the visual of two objects with empty space transforming to the two objects being together.
    - Pompidou 3y ago
      I love yout idea regarding language. I would check more deeply this idea. If so, I should examine 1) if language denote reality (maybe language is something else, like mutuallity of inner representation of reality, wich is different thant reality itself). If so, 2) if language can describe reality (maybe it will always miss some part). If so, 3), the sign (the written language) has to be able to fit to language (I don't know if we have the proof that language is 100% reductible to written language - think about written irony who let the reader sometime in absolutely indecidable choice, while in spoken situation it's clearly irony). It's a very vast subject ! Thank you for all the questions you raised in my mind.
andoando 3y ago
For simple programs you can also imagine just implementing a Turing machine where the symbols are colors.

boothby 3y ago

> What does an algorithm look like? Writing Piet by hand is a fun way to explore this. Sergei Lewis[1] and I[2] have each written tools to generate Piet code. Sergei's assembler makes much nicer looking code than my Piet backend. All you can really see from my compiler's output is that I'm really lazy and used a trampoline[3] [1] http://www.toothycat.net/wiki/wiki.pl?MoonShadow/Piet http://www.toothycat.net/wiki/wiki.pl?MoonShadow/Piet [2] https://github.com/boothby/repiet/ https://github.com/boothby/repiet/ [3] https://en.wikipedia.org/wiki/Trampoline_(computing) https://en.wikipedia.org/wiki/Trampoline_(computing)

andoando 3y ago

I have a belief that there is a 1 to 1 relationship between any algorithm, or even any mental conception, and a visual representation. I got this idea reading one of Steven's pink book, in which he says that any abstract word can be broken down into simpler words which in the end describe some spatial relationship. For example to "rekindle" is to "bring two things back together". It is easy similarly to say that a for loop is the mental conception of "one thing going over a bunch of other things", which has a visual representation like "100" -> "010" -> "001". I wonder then if its possible to create a language where one defines these constructs purely as visual transformations.

Pompidou 3y ago
> any abstract word can be broken down into simpler words which in the end describe some spatial relationship What if my computing model rely on quantum phenomenon ? Maybe some of them don't have good visual representation. But it's a guess. I don't know... Regarding the global idea of your comment: what about languages such as J it APL or BQN ? In such languages each character is an atomic part of algorithms (in j, two characters are used to implement dynamic IF for instance. One character only is used to define the dimensional depth of each part of algorithms). In this way, iversonnian languages can be seen as general and pure visual representation of algorithms once you consider each J or APL character as an arbitrary glyph.
- andoando 3y ago
  Well consider this: There is some set of words which describe quantum physics perfectly. If what I am saying is true (that human language maps to transformations of space-time) then this theoretical computer should be able to represent quantum phenomena. Im not familiar at all with APL or BQN but I think I understand what youre saying. I think the main point here isnt simply that you can represent programs visually, but that the visuals do exactly what you see. A "combine" instruction for example would literally be the visual of two objects with empty space transforming to the two objects being together.
  - Pompidou 3y ago
    I love yout idea regarding language. I would check more deeply this idea. If so, I should examine 1) if language denote reality (maybe language is something else, like mutuallity of inner representation of reality, wich is different thant reality itself). If so, 2) if language can describe reality (maybe it will always miss some part). If so, 3), the sign (the written language) has to be able to fit to language (I don't know if we have the proof that language is 100% reductible to written language - think about written irony who let the reader sometime in absolutely indecidable choice, while in spoken situation it's clearly irony). It's a very vast subject ! Thank you for all the questions you raised in my mind.

andoando 3y ago

For simple programs you can also imagine just implementing a Turing machine where the symbols are colors.