@stman Happy Birthday!

cc:
@theruran @vidak

@stman @theruran @vidak
That was interesting. I think you're very likely to find the same pattern right now anywhere. We've built massive systems dependent on continually accelerating growth, then starved the engines of growth by refusing to fund any pure research unless it shows very obvious signs of commercial application

The exception to the capitalist moratorium on pure research is the eternal hope that lightning can strike twice. When the rare academic manages to get funding for something actually novel, and that leads to commercial success, capital is willing to give him (because patriarchy) some access to resources he should have full control over by virtue of his labor. We're benefitting from this dynamic by way of Constructor Theory, from David Deutch, known for Deutch's Algorithm

@theruran @stman @vidak
This is in reference to your forum post, not your post here. This is the answer to "How does this extend to computer hardware?"

@theruran
I was just reviewing combinators to make a reply on krali. The short version is that all of Church encodings, particularly Booleans, follow from a minimal set of combinators. So it's possible to do practical computing with that as a starting point, assuming that mappings to silicon for at least one of those sets, e.g. S and K, can be made
@stman @vidak

capitalism is all about waste. ultimately the reason why is the reign of private property.

i do not see any reason why we should not treat computer resources the same as any other under capitalism.

so perhaps the observation that modern computers are still slow is due to the fact we are wasting significant amounts of their power.

this could be a good direction for the @permacomputer project.

web browsers, mandatory AI at the operating system level, computer languages with huge runtimes... all these are a target for ruthless criticism.

@stman @theruran @vidak
One of the reasons that human languages are ambiguous is that intent is, at best, inferred and, at worst, deliberately hidden. Famous examples include "ma fille" in French, which could be used to refer to one's wife, girlfriend, or daughter, and "love" in English, which might be used to more than 3 different ideas from biblical Greek

In highschool geometry, we did two-column proofs to communicate the idea of the relationship between the fact established and its justification, colloquially, "segment AB is congruent to segment CE because corresponding parts of congruent triangles are congruent," with "because" doing some heavy lifting. My assumption is that proof assistants, like coq, perpetuate that kind of structure, but that the actual convention may be different. In other programming languages, most notably Python, comment syntax is formalized so that external tools can automatically generate documentation

The practice of embedding the intent in the language, as opposed to using a macro language in comments, is what makes languages like Scheme and HTML 4 declarative. The practice of discovering intent from a social context is semiotics. I think one thing that we're hoping to accomplish is to make intents as discoverable as possible. To my mind, this requires 1) using declarative languages where possible, 2) using embedded declarations where necessary, and 3) interrogating the social context of legacy code for its intentions when unavoidable

The underlying principle here being that certifying software involves at least two steps, the two considered here being establishing the intent and assuring that the code performs that intent (and only that intent)

At @spritely we're building the future of decentralized networking tech (social networks and otherwise)! We just launched a supporter campaign and could really use your help! https://spritely.institute/donate/

I'm also going to talk about why Spritely is important & deserves your support! 🧵

things i wanna do:

permacomputer $0Ci4L M3D14Z

• fully update the @permacomputer project social media, which means:
  • updating the hackaday profile;
  • updating the master project page;
  • releasing a new statement on the mastodon profile.
  • (i think that covers it)

the permacomputer hardware itself

• build the first MC6847 video display 'warmup' computer.

  • i have found three projects based on 6502 which make use of the MC6847 and have published their schematics. so far so good. when the parts arrive i have to trust my breadboards and little wires are of a sufficient quality as not to bounce and ring, we're dealing with sub-1 MHz TTL-level digital signals so.... maybe my dodgy tools and components will be good enough.

• build the first permacomputer prototype

  • the next step after i have honed my practicals skills on a simpler task is this: construct a custom video display for the permacomputer, based on the information contained in don lancaster's cheap video cookbook or here if you wanna support the internet archive?
  • the reason for using custom logic for the display is the limited number of columns the MC6847--it is my hope to have 40 columns, not a mere 32. i wish i could say more, but the short story is this: using EEPROMs in place of discrete logic is the ticket to having a simple design.

the permacomputer software

• it is still my plan to get lisp running on the permacomputer (much thanks @50htz 🙏 🙏)
  • to this end see this disassembly of Acornsoft LISP for the BBC Micro. the beeb is a 6502 machine. it is my hope it can be easily ported to the permacomputer and enable a suitable operating system environment.
  • the problem is, this will make the permacomputer more like the beeb. it may not be a simple matter of reproducing the memory map of the beeb--it could require reproducing large parts of the beeb data pathways inside the permacomputer. watch this space.
  • i do not want to get too involved in writing a lisp for the permacomputer, so i am postponing this possibility until the very last moment. if i have to do everything myself, sigh, i will.

the overall 'permacomputing' approach

what is a permacomputer? what does a permacomputer do?

the ultimate goal of the project is to construct computers which will last a really long time, and still stay operational.

are you serious?

i largely regard this project to be a piece of art. if someone finds some practical use for what i am building, however, that would also make me overjoyed.

is this a kit project?

it is my pipe-dream to end up manufacturing permacomputer units at a kind of cottage-industry level, perhaps something for my worker's co-operative to do--friendly computers. i want to make these babies really, really cheap--if it's not cheap, it's not accessible to the poor.

why 80s TTL micro-electronics?

it should be taken as given that 80s micros are one of the last generations of computers that people could have a hope of comprehending and mastering every element of its specification, from hardware to the top of the software stack.

neocities is really one of the best things on the web, embodying so much of the spirit of what it means to use hypertext

makes me think if the solution to leaving web is to 'do a gemini' but for HTML--deliberately designing a protocol that is incompatible with the worst parts of HTML.

it's not altogether that difficult--just look at the simplicity of the gemini RFC

just wondering about it--from memory, most of the coding back during the geocities/myspace era was probably just copying and pasting little snippets here and there

i suppose the methodical way would be to step through the HTML5 standard and removing every element that contradicts our requirements

anyway, food for thought...

alright the 6502 thingo is going ahead.

i purchased the following

• 2 x MC6847 video display generators (keeping a spare on hand)
• 1 x T48 (TL866 compatible??) EEPROM programmer
• large amount of 74LS00 NANDs

i already have:

• 2 x 61256 32K SRAMs
• 1 x AT28C256 32K EEPROM
• 1 x ACIA
• 1 x VIA

-*-

links:

https://pmig96.wordpress.com/2020/07/15/mc6847-test-circuit/

https://github.com/adumont/hb6502?tab=readme-ov-file#memory-map

-*-

memory map:

-*-

honourable mention:

https://awsh.org/homebrew-6502-computer-part-1/

Screenshot of a gig work app asking how you plan to get to jobs. I'm pretty sure it is illegal for employers to ask if you have a car, but tech companies can because the law doesn't apply to them?

@stman I have to say that I've been too preoccupied with USPol to keep up internationally. This election here is really close and the outcome is really important for me, personally, so I'm watching polls and other domestic news constantly

Neither candidate will cut off weapons to Bibi, though, no matter how close to complete the Palestinian genocide becomes. Also, neither has commented on foreign policy outside of Europe and the Middle East

Odds are about 55% that the moderate Democratic Party will win the presidency and 45% that the fascist Republican Party will win. If the moderates win, the fascists should fall apart by midterm elections and the US will start to look like a liberal democracy with a real chance, at last, of socialists organizing into a major political force. If the fascists win, they'll do fascisms -- with all that entails