Fewer programmers interested in hardware?

I see far too many programmers these days who use macbooks and prebuilt desktops. They've never assembled a computer or installed a different OS. Why is this? I thought programmers were interested in tinkering with things. I feel more and more like I'm out of place when I talk about how I assemble and overclock computers and install GNU/Linux distros.

This is specialization and abstraction in progress. I can formally prove the theoretical mathematics behind computers and computation but don't ask me to have a look at your computer to check for viruses, I won't know how to do it.

>>1
Tinkerable computers are in short supply these days.

>>3
not if you know where to look

>>3,4
The move toward SoCs is a growing threat to hardware openness. It is now possible for a single chip vendor to provide a complete system with all code required to run user applications in a filesystem image; no source or datasheet need be provided to the end user. The sort of documentation and support that every systems programmer used to enjoy is now a privilege granted only to NDA'd employees of the largest hardware and software companies.

Ironically, it is the availability of a decent free and open source operating system (Linux) that makes this possible. Back in the bad old days, chip companies who had no hope of making an OS anyone would want were effectively forced to open up their specs so others could do that job for them. Today, they just add the bare minimum of binary blobs needed to get Linux to run and call it good. Want a manual? What manual; you need to place an order for a few million dollars worth of chips before their field engineers will so much as talk to you.

This is what happens when those idiots keep pushing their "learn to code" movement without forgetting that the majority of the population is basically computer-illiterate.

The same shit is happening with EEs, just look at tArduino and its followers. Suddenly everyone thinks they can build things when all they've done is buy some overpriced kiddie shit and read some instructions. You can teach a fucking monkey to push buttons, paste code, and connect things together and be poser EEsmakers and they won't know anything at all about how anything works.

>>2,5
Abstraction considered harmful. Companies are taking advantage of it to oppress the people.

>>1,6 (and everybody else)
I program to create worlds. I don't care about the hardware my world runs on, only that it runs my world.

On the Commodore 64, I programmed in Basic. I loved it because in my limited way, I could make abstract worlds. I didn't (and to be honest, deep down I still don't) care how it maps down to the level of the machine. If I had the books, I might have learned about Assembly, and built my layers of abstraction from there, but I would consider this work a necessary chore and not the core work of building worlds I'm interested in. I could give less than a shit about exploitable timing glitches in the sound chip or whatever tedious hardware hack. Ultimately, I want to program to make universes, not to piss around with chips. I still prefer to work from a high, interactive level. Knowing what I know now, I am angry that it shipped with Basic instead of Forth or Lisp, especially considering those were well known, well established languages by the time Commodore 64 came out.

>>6
And most people with CS degrees are also "computer-illiterate", it have nothing to do with any external agent.

Neo: Let me on.

Train man: No. I can't help you. Nobody can help you.

Neo: If you insist I'll have to resort to force.

Train man: No, you don't understand. I built this place. Here I'm G-d!

Train man punches Neo into the wall. Train man walks onto the train with the family.

Train man: You're going to stay here for a long time.

Neo: *cough, cough* *spits a little blood* -- If I recognize the sound correctly, you are using a Lenovo SFX blaster 680000?

Train man: Yes, why do you ask?

Neo: EEEEEEEEEEEEEEEEEEEEEEEEEEEEE ERRRRRRRRRRRRRRR EEEEEEEEEEEEEEEEEEEEE KSHKSHSKKKKKKKTTSTSKKTSSSSSSSSSSSKKKKKKKKK

A sharp metal pole protrudes from the floor and impales the train man

Train man: Hax..... *death rattle*

>>7
Nazis had a lisp for c64, although all identifiers are in 3rd reich sprache
https://www.c64-wiki.de/index.php/LISP_64

>>9
Now THAT is what the fuck I wanted to see in the sequels. Neo was supposed to be able to program the matrix as he sees fit (according to Morpheus' words in the first movie), so I was really looking forward to see him use programming to solve problems instead of punching and kicking like an idiot.

I wanted to see him invent funky weird shit like a program that converts a city block into a gingerbread house or whatever, and as a nod to his demoscene heritage, maybe program a blobby to capture the agents and turn them into cheese or something.
When he encounters the head AI guy, he compliments him in his hacks and they engage in a live programming duel, before realizing they're both awesome and being friends.

THE END

Instead the only character that got to do something like that was the "Merovingian" with his orgasm cake thing. The Merovingian rightly considered fighting to be beneath him, so he gets his bots to do it for him while he goes and programs orgasm cakes. I think the Merovingian has the right idea. He is the only one who comes out of the movies looking good.

Check out this Matrix:
0 1 5 7
2 6 4 2
1 2 8 5

>>11
That would be soooo coool. Much preferred to mediocre kung-fu superman versus less powerful brawling supermen. You've jailbroken reality and superman is what you do?

Instead the only character that got to do something like that was the "Merovingian" with his orgasm cake thing. The Merovingian rightly considered fighting to be beneath him, so he gets his bots to do it for him while he goes and programs orgasm cakes. I think the Merovingian has the right idea. He is the only one who comes out of the movies looking good.

fukin lol

>>12
Neo is trapped in the kernel! Noooooooooo

>>10
Neat. Recently I found a very good Forth implementation for the C64 as well:

https://github.com/jkotlinski/durexforth

As well as being really small and really fun to use, the assembly source for this is quite nice to read as well. Very neat implementation.

>>4
Where should I look?

>>10
LISP is an enemy of the JEWS which is why they suppressed it with fragmentation and endless revisions of SCHEME!

reverse polish notations

how can anyone take forth seriously

>>17
Whom are you quoting?

>>17
because it's effective? because they don't have tiny brains that can't think past syntax

RPN is best calculator

>>17
Their brains can only do one thing at a time.
When you say "sum the numbers 1 to 5" they think "push 1 onto the stack, push 2 onto the stack..."

Their brains can only do one symbol at a time.
When you say "dup dup swap mul add add swap mul mul" they think "k * ((z + (y + (x * x))) * w)"

>>22

"k * ((z + (y + (x * x))) * w)"

Stack monkeys can't even visualize an expression without a purely sequential evaluation order.

k * (z + y + x * x) * w
*/ k (+/ z y (*/ x x)) w

Right away you can see that x * x and z + y could be evaluated in parallel.
Algebraic laws can give you even more parallelism. Pack that shit into the SIMD registers.

can't believe you guys are hating on stacks

>>24
don't count me buddy

>>24
We got each other's backs.
We'll withstand the attacks.
Don't get those flash backs.
You may be tempted to grab your axe,
or send them an obscene fax,
and run miles on tracks,
but it's ok because we have stacks.
So relax, hit the sacks, and freebase some cracks!

>>24
Around stacks, never relax.

>>23
Whom are you quoting?

Modern C/C++ compilers are so optimized that’s almost impossible to a human to write a better code that generated by the compiler. That’s why the Vala language transpile to C with GObject and not directly in machine code.

>>29
Actually, compilers don't use SIMD opcodes to process arrays.

>>29
Bullshit.

>>30
Why should they? You can't assume that the processing for the array is safe for parallel processing. SIMD only makes sense for a certain type of processing like DSP work or computer graphics work.

dubs get

>>32
hence, compilers are shit at generating the best machine code

>>32
Yet most of the CPU time is spent doing DSP and graphics.

For example, neural nets are SIMD-friendly, so any serious NN package will have to include hand-crafted assembly core.

>>34,35
The point is that in order to use the special purpose computer (the SIMD computer), you need to use the proper tools to make it work. Using the constructs designed for general computing in a general CPU isn't going to make use of these special purpose instructions. Yes, you can craft a program by hand with SIMD assembly. You can also write a compiler that generates SIMD programs that'd take a much higher level approach than what humans would do by hand, but someone has to take the time to write such a compiler.

>>35

For example, neural nets are SIMD-friendly, so any serious NN package will have to include hand-crafted assembly core.

Name one.

>>37
http://papers.nips.cc/paper/718-a-massively-parallel-simd-processor-for-neural-network-and-machine-vision-applications.pdf

>>11

Instead the only character that got to do something like that was the "Merovingian"

whom?

>>38
That's a hardware architecture; the NN is written in Seeples.