The People's Permacomputer Project

vidak

The People’s Permacomputer Project

The People’s Permacomputer (henceforth known as "the people’s
computer", or “the permacomputer”) is a project dedicated to
resourcing the specification, and then construction of a
‘permacomputer’.

Who?

“The Committee”

Yes you can join the committee.

What is a permacomputer?

A permacomputer is a computer which attempts to embody the virtues of
permacomputing.

Foundationally, permacomputing itself is set of community practices
and traditions which shares a set of social and ecological values
inspired by the 70s land management and settlement design of
permaculture.

What is the point of this project?

The people’s permacomputer project is an attempt to physically
realise a permacomputer.

This will involve not just the production of an actual model
permacomputer, but also the development of a list of suggested social
and cultural practices around computing that will, it is hoped, assist
in the continued human practice of electronic computing.

There are many different dialectical approaches to making an
introduction to the people’s permacomputer. One thought experiment
that has proved especially popular and easy to grasp sums up the
mindset behind which we are functioning:

Industrial society has collapsed. All semiconductor fabrication has
ceased, society-wide electrification is no longer guaranteed. There
is no longer any internet. Computing as it was once known in the early
21st century is impossible. You need a computer for a task. What do
you do?

This project is a humble response to the challenge posed by the above
problem.

Adventures in the traditions of computation

We seem to take it for granted that a computer in everyone’s hand just
is democratic computing. Indeed, the ubiquity of contemporary
computation has been confused for ‘democracy’.

As quickly as we marched towards computing for the masses, we marched
just as swiftly away.

GNU

There do today however still exist flourishing movements which are
worthy of mention–although not exclusive in this honour, much of the
GNU movement is to be credited with any sanity being preserved in
present-day mass computation.

Hobbyist computing

The hobbyist computer movement of the 1970s was rich in ideas, and
courageous–sometimes breathtaking–in its efforts to allow the lay
person to realise their access to an electronic computer.

Need we speak of the heterogenous array of kit computers and their
attendant clubs and magazines? Some of mention are entire influential
computing platforms in their own right:

Altair 8800.
Apple I and II.
Commodore 64 and the VIC-20.

From minicomputers to microcomputers

One may even be able to recount the history of computing before its
entry into the mass consciousness. Computer architectures from (now
defunct) firms like DEC (Digital Equipment Corporation) still carry
enormous significance today.

Much of DEC’s fascinating and progressive work culture is imprinted on
the fruits of their labour. Two models of computer from DEC in
particular, the PDP-8 and the PDP-11, are steeped in the corporation’s
ethos: “do the right thing”.

Neither of these machines were of much relevance outside the academy
and industry, but they represent huge strides forward in human history
for the virtues that the people’s computer committee see as necessary
for permacomputing.

In particular, the full plans and maintenance manuals for each token
computer were accessible alongside each physical device:

http://www.bitsavers.org/pdf/dec/pdp8/

http://www.bitsavers.org/pdf/dec/pdp11/

When was the last time the entire structure of a modern smartphone was
exposed and made accessible to the user? Indeed, the devices we take
for granted today are deliberately obfuscated for the purpose of
unchecked economic profit.

Mutual exclusion

There are many influential projects which attempt to address the same
set of values driving the people’s permacomputer project. Some worthy
of note can be listed in no particular order:

Collapse OS.
uxn.
The RC2014 computer kit.
Ben Eater’s 6502 project video series.

All of these projects are concerned with some subset of the principles
the permacomputer project holds dear. Collapse OS is software that
aims to be system agnostic, and assumes the previous acquisition of
some supported hardware.

The uxn ecosystem is rich and continues to flourish. However,
while this project shares much with our own concerns for
sustainability and the long-term persistence of electronic computing,
we diverge from uxn on one fundamental point: we aim to
physically provide the community with complete systems that are
easily turned on and interrogated by the lay computer user.

That is not to say uxn does not also shine in such areas, but
uxn is (very consciously) an emulator. The people’s computer
project aims to quite literally place completed general purpose
computers into people’s possession.

While one horn of the dilemma may be expressed as "software without
dedicated hardware", the reverse can be said to be true of the
plerotha of projects similar to the RC2014, and Ben Eater’s breadboard
6502 project. In this case, it is "hardware without dedicated
software".

These two aspects of a general purpose computer–its own unique
physical construction, and its capacity to perform abstraction through
judicious programming–are usually siloed off from one-another
intellectually.

In this way, the people’s permacomputer is an attempt to blend, and
thoroughly combine two previously mutually exclusive set of practices
which, even if they were only able to survive the collapse in part,
are invaluable when applied correctly.

Permacomputing and heterogeneity

Our research was long and laborious. When the results of our
information gathering started to become clearer, we discovered that it
was not appropriate to offer up one model of permacomputer, but
several.

Mike Bauer’s DREAM 6800.
The DREAM’s ‘older brother’: a more powerful permacomputer able to
deal with more complex human demands.
Permacomputers constructed from salvage, such as e-waste. One such
proof of concept we imagine demonstrating is the collection and
repurposing of ESP32C3 microcontrollers from inside ’smart
lighting’.
Permacomputers derived from popular off-the-shelf contemporary
microcontroller kits, like the Arduino Uno or Mega.

Each of these four approaches manages to present a view of the
fundamentals of computing from a different perspective. It seems to
us, at least, that there are various competing demands that must be
balanced delicately in order to run a community organisation. There is
no ‘one size fits all’.

A mnemonic can be derived from the four computer models we humbly
offer up to solve your computing woes:

THE DREAM
THE OLDER BROTHER
THE SALVAGE
THE CHEAP + NASTY

to be continued…

theruran

vidak In this way, the people’s permacomputer is an attempt to blend, and thoroughly combine two previously mutually exclusive set of practices

This explicit scope of hardware and the software to use it is a very important distinction for the project, in my view. A lot of fun and learning can be had without ever touching the absurd complexity of modern computing.

Having a fully working, comprehensible computing system using silicon that has already been manufactured is an important response to a need that Big Tech will never, ever address. Bring back the days of getting a box of manuals with your new computer!

Because I assume the networking paradigm here would be sneakernet if not the old-school internet such as gopher, telnet and BBS with allowlisting. The cybersecurity implications are basically nil.

So I think the People’s Permacomputer is an important project and I am glad you all are working on it. I see this kind of project as complementary to our effort to design a new cyber-systems architecture that overcomes the systemic cybersecurity and complexity problems while being more convivial.

stman

Is it in the scope and context of the project [ Industrial society has collapse and mass IC manufacturing has ceased ] to also think about how to bootstrap low-tech ICs (Years 70 to 90′ IC manufacturing processes) re-manufacturing on DIY / Hackers way of doing things ?

vidak

Project phase two results

A ** Computer module. GITHUB REPO: USER AUDMONT

This proof-of-concept design is perfect for an ultra-simple and
easily-constructed computer.

This is the memory map:

Address - Component
$0000-$3FFF - RAM (16KB)
$4200-$4203 - ACIA
$6000-$600F - VIA
$8000-$FFFF - ROM (32KB)

The benefits of this design are many:

It uses a 6502, not the relatively exotic 6800
Will easily run FORTH and BASIC
Much easier and simpler memory addressing logic
Quite a luxurious amount of ROM, and this presents a good
opportunity to pack it full of interesting things, such as FORTH
words, BASIC variants with larger ROM footprints…
Possesses serial through the ACIA chip

B ** Video display processor.

The first phase of the project found that it was desirable to include
some form of visual output in the permacomputer.

It was decided not to include any display into the design of the
permacomputer itself, and instead require the user to provide their
own television.

Reasons for reaching this decision are that

it is expected that televisions will be possible to salvage in a
hypothetical societal collapse;
given that the video standard chosen by the project is analogue
composite video, this simple form of signal should have some hope
of being able to reverse engineered should cultural knowledge
around computing significantly degenerate.

The second phase builds on the results of the first. The work done in
the second phase concluded that it was prohibitively expensive and
cognitively overwhelming to expect the visual display capabilities of
the permacomputer to conform to video designs contemporary with the
1970s and 1980s.

MCU Video display.

The project found a solution to composite video output in small
microcontrollers. The idea is not to directly integrate video into the
8 bit computer design, but use a cheap microcontroller as a video
terminal, and have the computer use its ACIA chip to communicate with
it via serial.

Using microcontrollers does not contradict the spirit of
permacomputing. It radically reduces the chip count of the design,
allowing the project to proceed faster towards a
proof-of-concept. Using microcontrollers removes the need to sacrifice
system memory for VRAM. it also makes the permacomputer modular–if
one part of the system fails, you can still use the other half.

SCHEMATIC FOR BOTH PROJECTS

This consists of two modules and when used together produces a generic
ANSI terminal (also supporting graphics) using a TTL compatible serial
interface running at 115200 baud (can be changed in software) so can be
used for any computer project that has a serial I/O and needs a keyboard
and display.

This uses an ATmega88 (or 168 pr 328) for the keyboard and serial
buffer and an ATmega328 for the display processor.\ Since the display
is independent of the host (eg Z80) processor, the host has no
processing overhead so when connected to some other host.

The circuit is in two distinct parts, as shown below. If only a
display is needed then only the circuit to the left side of the line is
needed.

(NOT SHOWING DEFAULT FONT CONFIGURATION RESISTORS - SEE BELOW)

Note 1: Not all power supply pins are not shown. These must be
connected to the appropriate power rail.\
Note 2: Include R3 for NTSC display only, otherwise do not make any
connection between PD7 and ground.\
Note 3: Include R6 and also connect PB1 on the interface to ground
for 4-bit data ONLY. Do not connect any green wires when using the
4-bit interface.\
Note 4: Include R7 and also connect PB2 on the interface to ground
for two-wire ONLY. Do not connect any red or green wires.\
Note 5: If a specific display start-up configuration is always used,
the code can be altered (see below) and all configuration
resistors omitted.\
Note 6: Default font will be 80 char, single-height, bold. To change
it, pull the following display lines low with 10K resistors:\
Pin 6 (PD4) - 40 chars per line if pulled low via a 10K
resistor\
Pin 5 (PD3) - Double height if pulled low via a 10K resistor\
Pin 4 (PD2) - Non-bold if pulled low via a 10K resistor\

Monitor and keyboard controllers Controlled by low-cost ATMEL \“Mega\” processors.

Project 1: PAL or NTSC multi-mode text and graphics display with
I²C, 4 or 8 bit interface\
Project 2: PC keyboard-->serial and serial-->display interface\
Use both together to allow any project that uses serial I/O to have
a screen and keyboard***

Prototype.

The two parts of the above schematic (8 bit interface, PAL mode) are
constructed as shown here.

Note: this picture is for the schematic version 2.1. Minor wiring
changes for version 2.2

Processing and circuit description.

The serial/keyboard controller runs at 11.0592MHz (to coincide with the
115200 Baud serial interface) and the display processor runs at 16MHz.

The serialisation of the video data is done using a 74HCT166 parallel to
serial converter.

The sync and video signals are merged using a resistor network which
works fine for me and terminates in a standard 75 ohm video input on the
monitor.

The keyboard used is a standard \“PC\” keyboard with a PS/2 connector
(NOT USB). Caps lock and num lock works and it updates the keyboard LEDs
as needed. Non-ASCII keys return codes, but this isn\'t finalised yet
and is easily changed in the code.

ATmega88/168/328 processor (Serial and keyboard I/O).

This buffers the serial input and stores it in a circular buffer 800
chars in size. The RTS is cleared when less than 32 character spaces
remain in the buffer and is set when there are less than 32 char spaces
used. This is to allow overrun from the sending system without losing
characters.\

When a character or command is to be sent to the display processor, the
code is placed on the output port pins.\
Some ANSI commands are interpreted and converted to the relevant video
processor commands.

ATmega328 processor (video).

This has 2K RAM so can accommodate an 80×25 char display with some
bytes remaining for workspace. The main processing is transferring
data from the memory to the output serial shifter, and producing the
HSYNC and VSYNC signals. During the inactive time of the display, the
input port/status bits can be read and the incoming data interpreted
either as a command or as a character that is inserted into the
display. Scrolling and clearing of the screen is handled within this
processor (very quickly!).