@Shamar @natecull here’s an article on just how much of a struggle it is to actually teach kids to read

https://www.apmreports.org/episode/2019/08/22/whats-wrong-how-schools-teach-reading

and here’s roger penrose on the question of whether math is invented or discovered

https://youtu.be/ujvS2K06dg4

@zensaiyuki @Shamar @natecull god. i learned to read the old phonetic way, and had to suffer through a school system switching to the "whole reading" bullshit that produced a bunch of barely functional illiterates.

have pretty much binned everything i learned in school and went back to whatever they did in the 1920's.

@icedquinn @natecull @Shamar some of this stuff came up in the recent discussion of an a possible ancient prehistoric civilization (the silurian theory). i was put off by the particularly dingbatty tone of one youtube video about it so I looked up the definition of “civilization” that archeologists use. in sum, you need a city, a government, a class system and specialisation, all only really made possible by a writing system of some kind, for receipts and spreadsheets.

@icedquinn @natecull @Shamar you can debate whether any of that stuff has in fact been “good” but there’s no dispute that reading and writing is a technology.

@zensaiyuki @natecull @Shamar writing systems are kinda sorta not needed but they help. there's that one tribe where currency are big round rocks and you just have to remember the entire oral history of each rocks trades. they are unironically using a blockchain since years

it... helps though.

@icedquinn @natecull @Shamar it’s a bit of a stretch, but you could call those rocks a kind of writing system. it’s about a step away from a quipu, which *is* a writing system. of a sort.

@zensaiyuki @natecull @Shamar the history of the rocks and current owner is carried in oral history. the rock itself bears no data; it just looks like a really big yen made out of stone.

@icedquinn @natecull @Shamar it’s presumably uniquely identifiable. that is, technically, data. but, it’s really arguing semantics.

@icedquinn @natecull @Shamar and- to be clear, it’s possible there were writing systems older than 6000 years. the trouble being the complete absence of evidence for them. which could be because the materials have been obliterated by time. paper and wood doesn’t last very long without climate control.

@zensaiyuki @natecull @Shamar pretty much. i think taleb mentions this in one of this books. people thought some old society had no romance novels and only ever wrote serious business, because the business records were in stone and the pulp was on paper. until someone actually did find a paper book that somehow survived.

@zensaiyuki @Shamar @natecull not to mention part of taking over societies is to explicitly iconoclast their culture. we know gaelic is a language for example, but good luck getting most of the stories because they were deleted or modified by christians.

@icedquinn @natecull @Shamar back to smalltalk- Alan Kay has a talk “doing with pictures makes symbols” which is basically his argument against the very concept of a “programming language”, in the way it constrains our thinking. symbols are the important thing, not writing.

@icedquinn @natecull @Shamar Bret Victor tries to continue that train of thought in “Media for Thinking the Unthinkable” http://worrydream.com/MediaForThinkingTheUnthinkable/

@zensaiyuki @natecull @Shamar i am somewhat horrified the depths you have all gone in to smalltalk worshipping in this thread. actually i didn't know it was a smalltalk worship thread

i'm so proud.

@zensaiyuki @Shamar @natecull I've only ever known or worked with smalltalk-80 and its mutant cousin objective-c, so "Blobcat shitpost: 'aaaaaaaaaaaaaaaaaaa'; selfDestruct." is all i've known

@Shamar @natecull @zensaiyuki although wondering how to fix cog's derpy culture is what lead me to learning how the Shenandoah GC works.

@icedquinn @natecull @Shamar i know some of those words!

@zensaiyuki @natecull @Shamar imagine a high school nerd deciding to write a compiler without even a copy of the smalltalk blue book.

it took a long time to work up terry davis powers.

@zensaiyuki @Shamar @natecull well i learned from nim and CLOS that if you use binary dispatch on object identities, this is weirdly faster than some dispatch tables. but this requires redoing your whole VM so instead of looking up objects and trying to resolve selectors to them, you look up selectors and resolve object types to it.

for whatever reason (i guess because the O(log(n)) thing?) this method of dispatch is faster than vtables. so my tendency to want to fiddle with shit also got in the way.

@Shamar @natecull @zensaiyuki it also means that when new objects have similar selectors you have to modify the selector table, so you are basically doing some forth-esque rewriting of dictionary words at runtime.

@icedquinn @natecull @Shamar that setup almost suggests an accidental polymorphism. a runtime indexed ducktyping table. interesting!

@icedquinn @natecull @Shamar who wants this message! free to anyone who can understand it!!

@zensaiyuki @natecull @Shamar yes, well CLOS basically does it.

(defmethod aaaah) establishes that screaming exists, and then you define individual screams for particular clos object types. the dispatch for screaming then becomes a binary search over pointers to the object's metadata.

there were also some deep black magics where through the use of type specifiers and declaims, you could even make certain procs that simply bypassed dispatching entirely because it already knew what type the function accepted in advance.

i think c++ 'tries' to do this but is a lot more restricted about it

@icedquinn @natecull @Shamar i am an affete dabbler. i am more interested in ideas than actually doing anything with anything. the concept of older versions of smalltalk being like forth is fascinating . if you go to the paper linked in nate’s original post, it links to live web based emulators of each version of smalltalk to try out.

@zensaiyuki @natecull @Shamar i have made a couple attempts to interpret st-80 although most of those times i lacked big brain energy.

@icedquinn @natecull @Shamar i lack the big brain energy too- but i did find the interesting project SOM (smalltalk object machine) when researching an adjacent thread nate is in. they were talking about how modern language research lacks backtracking and tail call optimisation because the VM’s/ecosystems/runtimes they’re built on lack those features, and thus limit the big risky ideas new languages are willing to try. i thought “big risky ideas vm” would be an interesting project.

@zensaiyuki

1. How old is painting? The point is that asyncronous comunication that these tools enabled shaped us as a species.

https://en.m.wikipedia.org/wiki/Painting#/media/File%3ALubang_Jeriji_Sal%C3%A9h_cave_painting_of_Bull.jpg

Yet you are correct that universal literacy is a recent thing. But despite being a slow process, evolution preserve only those traits that become universal.

2. The truthness of #Pithagoran Theorem is an interesting objection, but I'd argue that it supports my point: have you ever seen a triangle in nature? have you ever seen a square or a circle? They do NOT exists.

#Math IS discovered, actually.

But it's always a discovery of humans' minds' traits and behaviors. We see math outside us but it's just a projection of our minds on our perceptions. It's like looking through invisible glasses that shape what we see. By discovering math, we just discover features of these glasses.

And math is so much about humans that a theorem is not proved until all humans can understand it.

@natecull
@icedquinn

@Shamar @natecull @icedquinn triangles and squares in nature? lots of em. and I don’t think they’re a cultural hallucination.

the asynchrony of painting? ok fine, no objection to that. but it has no bearing on the unnaturalness of reading and writing. staying mindful of the original topic, my point is that “programming languagess” based on text unecessarily constrains our imaginations for what computers are capable of. categorising paintings as valid communication is agreeing with me.

@zensaiyuki

I don't see a single straight line in the images you shared. Even when they look straight (as with some cristals), they are not. And the truthness of Pythagorean theorem breaks on non euclidean spaces and below the Plank constant.

You see straight lines, triangles and so on just because you are abstracting on your perceptions. Which is fine, fun and useful, but abstractions are just in your mind. Since we are both humans, I can see them too. But I'm simply aware they are just constructs of our minds.
Different senses (perceprions) and brains and thus different evolutions would lead to completely different maths.

___

As for the constraints that language pose on programming, yes I could agree that a different expression system would lead to different programs.

I play piano.

I guess that if I designed a programming system with a piano keyboard as the only possible input system, the programs I could invent would be very different from the ones I code in the several programming language I know.

BUT, the choice of language and writing to express programming was not driven by the available input devices, but by the available humans for whom such devices were built.

It's correlation, not causation.

Writing, so far, is the most ptecise and efficient way humans invented to comunicate asyncronously, as we do with the users when we write software that they run (or even just interact with).

@natecull @icedquinn

@Shamar @natecull @icedquinn it’s fun to think about universes with differerent rules, and alien perceptual systems. i think though, that if you have gravity, light, a body that is capable of movement in an environment, a need to eat and reproduce, a lot of the problems that alien perception would need to solve would be the same.

as for writing being precise and efficient? maybe for some things, but not all. Writing can’t capture tacit knowledge.

@zensaiyuki

Actually you don't need to think about alternative universes to meet non euclidean geometry: the Pythagorean theorem doesn't work over the Earth surface.

And you don't need to involve aliens: you can just think about dolphins, blind worms, flies or ants: different perception systems would lead to entirely different abstractions.

For sure writing can't capture tacit knowledge, but the programmed computers cannot either!

I wrote about this before https://medium.com/@giacomo_59737/the-delusions-of-neural-networks-f7085d47edb6 and https://medium.com/@giacomo_59737/yet-another-definition-of-intelligence-9bbaaa73086d

So if you want a dumb computer to do what you want, you can't use a language that leverage tacit knowledge that it lacks.

@natecull @icedquinn