TIL in linguistics : English is called an 'analytic' language ( few affixes ) whereas Malay is called a 'synthetic' language ( many affixes ).
Structure of conscious thought. Sensory data is bound to space and time. What are the smallest units of space we can model with machinery?
2D : pixels : circles, 5-adjacent; 3D : voxels : spheres, 11-adjacent; elegant, but unconventional. Return to this later! Regress to the Cartesian mode : cubes, 26-adjacent : wasteful corners : N-cubes of voxels :
This is an early stage analysis, but it is quite encouraging. It matches the intuition that the weight of thought is not as great as many people make it out to be.
I generally put it to designers of education systems : the survey of "what jobs exists, how much they are paid, and what is needed to access them" should be available to all 14-year-olds. It's the best time to start planning.
And the incumbent nature of the political economy is that, the absence of such survey data for 14-yos, is how we manage to keep people in lanes for the rest of their lives. :)
The 14-yos who know that the fresh grad pay range in KL is 1.5-20k MYR/mo are not doing what the 14-yos who think it's 1.5-5k are doing. And the expectation diverges for say a 30+yo where some make 1.5k, and some make 400k!
So the strategy of picking education paths depends on education goals, and the goals depend on ... some sort of preliminary life planning survey which is generally unavailable to the masses.
Humour : is simply a pattern of motivation, in immune systems.
Structurally it consists of no more than threat and mitigation. Why is this valuable? Because a system that practices, is a system that learns.
Therein we find value in humour, as a coefficient of education, and education as the process of civilisation's development of immunity.
Programming languages are obfuscating, in the same way that degrees and job descriptions are obfuscating. The political economy of this is intriguing.
In all activities, humans are doing very similar things. But for the purpose of "laning" or "putting people into manageable boxes" we often create barriers of language which establish structures of containment, and then we gatekeep the containers.
Anyway ... programming languages :
... and generally it is not easy to find a single tool that caters to everyone's needs.
But ultimately everyone's needs are the same, and so a tool should exist which serves everyone, with an opt-in for any specifics.
As always, the terminological conventions of math, have nothing to do with math. I almost never bother to remember which is left and right.
1.
The convention in matrix ordering is (y,x) which is transverse to Cartesian conventions of (x,y). Definitely learnt this in school, but never used matrices for years, so didn't remember it. Now that memory-layouts are a thing, and I'm building a programming language for knowledge management, I shall need to bear it in mind.
2.
This terminology of matrix ordering carries into computer programming conventions, whereby
handle-index-index is interpreted as handle-y-x i.e. handle-rows-columns
3.
Subsequently, the terms row-major order and column-major order can be interpreted only under this assumption.
row-major ordering is where the data for handle-*rows*-columns is stored together in memory, as is the case for languages : C / C++
whereas,
column-major ordering is where the data for handle-rows-*columns* is stored together in memory, as is the case for languages : FORTRAN, Julia, R
3b.
As a further consideration, the use of Iliffe vectors and linked-lists, as in the simplest data structures of many high-level programming languages like Erlang, Pyhon, and Javascript is not relevant by this concept.
TIL terminology : the indexing of dimensions and tensors, is different.
So
Ya?
MathWorld : ""An nth-rank tensor in m-dimensional space is a mathematical object that has n indices and m^n components and obeys certain transformation rules....Tensors are generalizations of scalars (that have no indices), vectors (that have exactly one index), and matrices (that have exactly two indices) to an arbitrary number of indices."" - via Reddit
There are a few issues on the table.
Every programming language has 3 main layers.
So why do I need to develop a new language? Well, I don't but it is convenient to do so.
Oh well. What a chore.
In the beginning, business was wrought from chaos.
Then came accountants, and there was order upon the land.
In a season of low margins, they were reborn as management engineers.
In another season, as consultants and bankers.
In yet another season, reborn as big data analytics experts.
In YET another season, as data scientists.
Today all the kids are running around discussing degree programs with "data" in them.
Hello, did you forget the "science"?
Solved a big design problem today :
"a<*A, b</*B, C*/>c, D*>d" replaces "a∈A, b∉B, C∌c, D∋d"
( Works best for full-height-* fonts. )
1.
So in maths, we write "a∈A, A∋a" to mean "a is an element of the collection A", and "a∉A, A∌a" to mean "a is NOT an element of the collection A".
This is basically a sexy round version of the Greek letter, epsilon, and it is shorthand for "est", meaning "the entire work". This is attributed to Peano.
2.
Our common ASCII character set has no such equivalent. It would be UX-wise ridiculous to write 'e' everywhere. So for the past couple of weeks I've been using "<e </e e/> e>" in place of "∈ ∉ ∌ ∋", which I thought was pretty smart.
That was until I started using this for Type Hints/ Annotations/ Signifiers : where writing something like "variablename<eTypeA" has a poor UX, due to the adjacency of the latin [e] and the [A-Z] charset for the first-letter of type names. This is no problem for a machine lexer, but humans are less clear in thought. And it also depends on the limited charset for type name first-letters, so phooey.
Oh yes, and using '∈' for Type Hints instead of the paamayim nekudotayim 'value::type' ( as in Haskell), 'value:type' ( as in Python, OCaml ), 'type<value>' ( C++ ) is, indeed part of the solution.
3.
All that being said, the revision is a work in progress.
Brief points from chat on Friday :
