Interesting technology, systems, and concepts
editTechnological liberty
edit- copyleft
- Librenotgratis.org (not yet launched)
Software
edit- libre software
- the GNU Project
- GNU Affero General Public Licence (AGPL)
- the GNU+Linux operating system
- Respects Your Freedom (RYF)
Hardware
edit- open-source hardware (OSH)
- Arduino open source PCB
- BeagleBone Black and other BeagleBoards
- OpenCores open source chip designs
- OpenRISC
- field-programmable gate arrays (FPGAs)
- software-defined radio (SDR)
- Novena
- Open Lunchbox
Resilience, decentralisation, reliability, security
edit- Dunbar's number
- multilateral exchange
The Ripple monetary system– appears to have become corrupt (circa 2010), but I still believe in the original decentralised concept of chained IOUs, with no tokens or ‘coins’ flowing through a network.- barty economics
- information-theoretically secure cryptography
- one-time pad encryption
- one-time authenticators (e.g. Poly1305)
- cardinal voting
- delegative democracy
- web of trust (WoT)
- peer-to-peer (P2P)
- distributed revision control
- mesh networking
- Stellar Consensus Protocol (SCP)
Efficient, renewable, sustainable transportation
editThis subsection is under review. Considering that the price of coal skyrocketted as a direct result of James Watt inventing a more efficient steam engine, and that this is far from an isolated example of inventions increasing the temptation of natural resources, I've lost confidence in the assumption that more efficient technology is what's needed to solve the world's sustainability crisis. It seems counterintuitive at first, but the market dynamic of efficiency increasing resource value makes sense – if a tonne of coal can do twice as much work then industry may perceive it as twice the value. Markets that formerly deemed coal too dear, such as mills, started to see it as economical, increasing demand, driving up the price, and thus motivating more agressive extraction – the onset of the Industrial Revolution. It seems to me that more efficient technology is simply more addictive technology, giving more economic justification for a particular amount of environmental destruction, and thus inventions of technological efficiency could actually just be enabling humanity to destroy Earth's ecosystems even more efficiently!! However, it's clear that inefficiency is bad too. There's a moral dilemma here, and I think that the emphasis has to be on sustainability from the outset, not on efficiency as an afterthought. |
- ratchet-based infinitely variable transmission (IVT)
- regenerative braking
- ‘brachtrain’ – a train in a tunnel that follows a brachistochrone curve (a cycloid with cusps pointing upwards), optionally with an isograv (horizontal) segment where depth is constrained. (Minimum horizontal length is π×depth; lengthtunnel is only (4 - π)×depth longer than lengthhorizontal; vmax = √(2g×depth); core journey time is (π×depth + lengthhorizontal)/√(2g×depth); 0 < [perceived acceleration (entirely along the train's vertical axis)] ≤ 2g.)
- ‘machtrain’ – a vacuum brachtrain whose tunnel is at least 6km deep, thus reaching mach 1. (In stark contrast to HS2: lengthhorizontal = ~160km (as-the-crow-flies between Birmingham New Street and London Euston/St. Pancras but with negligible environmental destruction); lengthtunnel = ~166km; vmax = ~343m/s = ~1235k/h = ~vsound; core journey time would be ~521s = 8min + ~41s; about 2×55s of abnormal gravity. Yet almost no operational power is needed! Tunnels would also be a powerful geothermal source with temperatures over 150°C.)
- ‘brachtrain’ – a train in a tunnel that follows a brachistochrone curve (a cycloid with cusps pointing upwards), optionally with an isograv (horizontal) segment where depth is constrained. (Minimum horizontal length is π×depth; lengthtunnel is only (4 - π)×depth longer than lengthhorizontal; vmax = √(2g×depth); core journey time is (π×depth + lengthhorizontal)/√(2g×depth); 0 < [perceived acceleration (entirely along the train's vertical axis)] ≤ 2g.)
- biofuel
- green wave
non-rocket spacelaunch– I nolonger believe that any sort of spacelaunch can be sustainable.
Miscellaneous/unsorted
edit- Stream Control Transmission Protocol (SCTP)
Haskell purely functional language– Haskell was my favourite language between circa 2012 and mid 2020, but I never managed to master it and eventually I realised that most of the performance issues were due to inefficient primatives rather than my own skill. Switching to GForth as my go-to language, I quickly achieved 100× greater CPU efficiency on my GForth rewrites of bits of Haskell code, despite being a novice in GForth and doing so in much less time than it took me to write the original code in Haskell. It's fair to say that I gave Haskell a good old try, but now my Haskell days are over.program derivation and proof-carrying code– How do you verify the verifiers? I've come to realise that I would never be able to verify the kind of complexity in an automated theorem-prover. Moreover, I've come to realise that I cannot even verify a mathematical proof as complex as that of Poly1305, due to Gödel's Incompleteness Theorems. I would basically have to be so competent in every field of mathematics and number theory that the proof touches as to be certain to notice any error in the proof, and such certain competence is basically impossible due to the nature of the incompleteness theorems. So essentially, while I can stare at the proof for weeks and find nothing, I'm still taking the word of those sufficiently competent in number theory that the proof is correct. Therefore, I've turned more strictly to minimalism and ‘implementation-simplicity’ as a viable alternative to complex verification techniques.- Low Level Virtual Machine (LLVM)
- Kernel-based Virtual Machine
- Chromium (computer graphics)
- Gallium3D
real-time ray tracing (RTRT)– Energy-intensive and contrary to minimalism.- Wicki-Hayden isomorphic musical keyboard layout
- Hexiano isomorphic musical instrument for Android
- lossless compression – Reduced technological lock-in because lossless formats can be compared objectively and conversions between competing formats do not cause further losses, as with lossy formats.
- Opus libre speech compression format
- WebM/VP8 encoded HTML5 Video
- Unlicensed Mobile Access (UMA)
- Ubuntu for Android
Gloss Haskell graphics library– Defunct.- IEEE 1394 (AKA FireWire)
- Xen bare-metal hypervisor
- heterogeneous multiprocessing (e.g. ARM big.LITTLE with GTS)
- ARMv8-A (Although ARM is proprietary hardware, I consider it more open than Intel – the lesser of the evils of the mass-produced chip architectures. Indeed most open-source hardware is based on ARM.)
- kneel chair
Dream devices
editAny dream device would use entirely libre technology, including all hardware components. This goes far beyond RYF which only addresses software; it would be the ultimate in technological liberty. Furthermore:
- there must not be any planned obsolescence – devices should be expected to last at least 10 years of frequent use and be repairable/serviceable;
- storage must always be removable;
- batteries must always be removable and separately chargeable (preferably hot-swappable);
- there must always be at least one physical port for a wired data or power connection;
- the charger/dock port is the same connector and is symmetric power and data (bidirectional), akin to PoE;
- port extension over the symmetric data connection to port-extending power supply;
- the device must be capable of natively running code written in a purely-functional language (e.g. Haskell);
- the operating system's packages must be manageable by a purely-functional package manager (e.g. Nix);
- general-purpose devices (i.e. not dedicated/embedded) must be compatible with a fully libre GNU+Linux operating system (e.g. NixOS with
allowUnfree = false;
); - portable devices must be fanless/passively cooled (constraint on power consumption) and stationary devices should at least be operable without a fan (What about heat pipes to chassis / heat spreader?);
- handheld devices should be at least 2 metre drop-resistant and at least 3 metre water-resistant;
- screens should be efficient and harness ambient light energy (dark themes use less power) rather than be hindered by it (i.e. sunlight-readable);
- handheld screens should have at least 10-pointer multitouch and inductive stylus support with an inductive stylus slot;
- radio interfaces should be SDR and support receive-only services such as AM and FM broadcasting;
- E-Loran and, in general, ‘terrestrial positioning system’ – like GPS but broadcast by stationary, nonsatellite transmitters such as the cellular network, media broadcasting transmitters, Wi-Fi access points, etc. (stations broadcast their precise location and time; devices use multilateration to determine their precise location and time, and if devices use an antenna array then orientation too (i.e. position and time); subdecimetre precision);
- Wi-Fi 802.11n, Bluetooth LE, etc.;
- mesh networking;
- radio transmission should have a physical power switch (for radio silence, a.k.a. ‘flight mode’);
- cameras should be stereoscopic and support at least 120f/s video capture at 1920×1080;
- camera and microphone should have a physical power switch;
- hardware compression should support lossless formats such as Dirac and FLAC;
- chassis should have an anchor point, particularly for handheld devices;
- display resolution should not exceed 75px/° at typical viewing distance (anything greater is unnecessary and wasteful);
- displays should be stereoscopic or volumetric;
- video out and video in over a symmetric data connection (e.g. via Ethernet);
- infrared and Li-Fi transceiver;
- screens should have infrared beacons built-into the bezel for use with a Wiimote.
Dream smartphone
editMy dream smartphone would have the following features:
- (common features);
- about 80mm×160mm×15mm;
- stereoscopic 2×8.3Mpx rear camera (à la LG Optimus 3D and HTC Evo 3D) and 2.1Mpx front camera;
- volumetric display with at least 2 views (e.g. left eye, right eye, left peripheral, right peripheral);
- display resolution exactly 1920×1080 per view, 24b/px, 60f/s;
- stereophonic speakers (à la HTC One);
- FM broadcasting receiver;
- portable desktop mode (à la Ubuntu for Android);
- 130mm ≤ display diagonal ≤ 160mm;
- anchor point and ports at bottom;
- ability to reduce the rate of charge or suspend charging altogether;
- active dual-SIM;
- UMA (à la Samsung Galaxy S II);
- ports: MicroUSB (charge, OTG, MHL), MicroSD, infrared/Li-Fi transceiver, 3.5mm audio out jack;
- density < densitywater (998.2071kg/m³ at 20°C);
- mass ≤ 170g (80mm×160mm×15mm×998.2071kg/m³ = ~191.7g).
Dream hybrid laptop–tablet
editI currently have 2 ThinkPad X60 Tablets, which is about as close as I've found so far to what I want. However, I've since seen devices that achieve the same hybrid laptop–tablet roll by having the screen detach entirely as a standalone tablet. This is more flexible and allows the tablet mode to be much less bulky. I prefer to think of this as actually having a screenless laptop with a tablet which can be used as the laptop screen, i.e. that both parts are autonomous with their own CPU and operating system, with software integration. For best flexibility, the tablet should be able to dock in both landscape and portrait, the portrait mode being especially useful for reading documents and such.
Dockable tablet
edit- (common features);
- about 396mm×240mm×10mm (with 20mm bezel around screen);
- screen about 356mm×200mm, 408mm diagonal (~16.1 inch);
- dockable in landscape and portrait;
- can connect to power supply directly, via the unused dock port;
- at least 2GiB of RAM, and an SSD of at least 128GiB;
- stereophonic speakers;
- ports: 2× symmetric power and data / PoE (landscape dock, portrait dock), 2×USB (on other 2 edges), MicroUSB (charge and OTG), SD card slot, infrared/Li-Fi transceiver, 3.5mm audio out jack;
- density < densitywater (998.2071kg/m³ at 20°C);
- mass ≤ 800g ((200mm×16/9 + 40mm)×240mm×10mm×998.2071kg/m³ = ~947.6g).
The capacitative sensing should have accurate position and pressure sensitivity with at least 10 multitouch pointers for best enjoyment with Hexiano.
Screenless laptop dock
edit- (common features);
- about 396mm×240mm×15mm;
- a pointing stick on a backlit keyboard with a standard layout, and no touchpad;
- swappable batteries at front side (for centre of mass stability; less dense components nearer hinge);
- at least 2GiB of RAM, and at least 256GiB of storage;
- FPGA;
- software-defined radio supporting AM/FM and DVB-T;
- stereophonic speakers;
- ports: 3× symmetric power and data / PoE (left, right, dock), 2×USB (left, right), MicroUSB (charge and OTG), SD card slot, infrared/Li-Fi transceiver, 3.5mm audio in and out jacks;
- density < densitywater (998.2071kg/m³ at 20°C);
- mass ≤ 1.2kg ((200mm×16/9 + 40mm)×240mm×15mm×998.2071kg/m³ = ~1421g).
Dream desktop / router / home server / DVR
edit- (common features);
- ultracapacitors for temporary power failure or swapping of power supply;
- at least 4GiB of RAM, and at least 1TiB of storage;
- 2 slot-load optical drives supporting CD-R, DVD±R, and BD-R, one of which is physically readonly;
- FPGA;
- software-defined radio supporting AM/FM and DVB-T;
- ports: 4×PoE Ethernet, 4×USB, 1×FireWire, SD card slot, infrared/Li-Fi transceiver, 3.5mm audio in and out jacks, digital audio out;
- …
Port-extending power adaptor
editThese days devices seem to be ditching the ports, but until devices are efficient enough to run and charge off a photovoltaic backing, they will still have a power port. Having one cable plugged in isn't a problem, so why not additionally use this same cable for a network connection and a port replicator?
For example, the power adaptor could have an Ethernet port, HomePlug connectivity, 4 USB ports, 1 FireWire port, VGA, DVI, HDMI, and digital audio. Wouldn't that be just excellent!
Favourite GNU+Linux distributions / libre operating systems
edit- NixOS – Operating system using the purely-functional Nix package manager.
- CyanogenMod – Distribution of Android.
- Replicant – Android replacement aiming to be fully libre.
- OpenWrt – Largely libre firmware replacement for routers.
Arch LinuxUbuntu