Myth of Being Digital
Exploring the Line, the Machine, and the Artificial Promise of Encoded Knowledge
A thesis exploring the age-old pursuit of knowledge in the electronic digital era.Chapter Bullets Explained
○●○●
Throughout the thesis you will notice four dots ○●○● at the start of each chapter. This is a digital binary encoding of the chapter number. If you wish you can play with the encoding yourself using the tool below.
How Fast is Your Machine?
Ada Lovelace, became the worlds first programmer when she wrote instructions for the Analytical Engine, created by Charles Babbage. Lovelace must have had a world of patience since the machine took three minutes to ‘multiply two twenty-digit numbers’.
Below is a simple program created for the purposes of this thesis that lets you compare your digital machine to the Analytical Engine.
Note: This code avoids FPU double or long data-type operations because 64-bit systems cannot hold full twenty-digit values, let alone fourty-digit answers. Instead, the software arbitrary BigInt data-type is used to achieve the same perfect precision as the Analytical Machine. This test also compares your machine to the PDP-1 and OLPC.
What is Digital Data?
At the lowest level digital information is stored in a 0 and 1 binary format. In this encoded symbolic state there is no indication of media: Is this audio, image, text?
Message, and meaning, are effectively destroyed in the process of digitization. In its stored state digital represents everything and yet nothing at all. And so, a machine does not read or recall information, because no message exists. The machine must fabricate a message every time data is accessed.
The inquisitive reader can us the tool below to explore their own media. You can upload anything you want.
Drag & drop file
or click to browse files
Maximum file size: 100MB
The Ethics of Observation and the Violence of Data
θ
Revolution is without Resolution
Qubits are like classical bits in that they are both conceptual units of information. Just as a bit can be physically represented by any system with two distinct states—like a transistor’s switch, a finger being up or down, or the position of a bead on an abacus—a qubit can be represented by any physical system that possesses the quantum property of superposition, such as an electron’s spin or a photon’s polarization.
What’s interesting is how this connects to the nature of physical properties we perceive, such as rotation, which are continuous—or analogue—in the sense that they have no resolution. This quantum property is traditionally visualized using the Bloch sphere I have provided below. The state of a qubit, like our example photon, can point to an infinite number of possible locations across the surface of this sphere. The surface is a manifold, meaning perfectly smooth with no discrete steps between one location and another. Much like the hand-drawn line discussed in the thesis.
When humans measure, quantum phenomena, or the continuum, collapses yielding a binary result of 0 or 1. The seemingly simple act of measurement, simply observing the phenomena, is a violent event in the quantum world. It forces the qubit to give up its superposition and “snap” to a definite classical binary answer.
This thesis uses quantum principles to argue that ‘data’ does not simply exist; data is created in the act of gathering. This emphasizes that data not only filters according to bias but that measurement itself is a destructive act, destroying information, and with it possibilities, carried in the analogue. The same applies to measuring a cultures embodied beliefs, collecting data on a culture demands those affected effectively violate their rituals which cannot be measured- and so observation too is a profound act of violence.
It is important that the reader keep in mind that perception of that act matters. For example, a doctor measuring a patient’s vital signs also collapses a continuum (blood pressure, heart rate) into discrete numbers, but the ethical context (healing vs. extraction) matters.
Bit Rot
How transistors work, and when they don’t
Left to the vagaries of environmental conditions, without maintenance or power, transistors eventually lose the charge necessary to maintain a 0 bit value (by default, without charge a transistor is 1). This non-volatile memory, called NAND (“Not-AND”) flash memory, is found in most storage devices today. Memory registers store this data by trapping electrical charge in a series of transistors (a 64-bit system would have memory registers that hold 64-bits).
Cells gradually leak charge, similar to a battery. This accelerates when heat is introduced, and when the wear & tear of ‘Program/Erase cycles’ increase. If too much charge is lost, the bits become ambiguous, difficult to distinguish, and the data is then consider corrupt. This is what is referred to as bit rot. When enough charge has leaked the transistor can no longer perform its main function, to hold a value (0 state in the case of Single Level Cell transistors). When this happens data is considered corrupt. To prevent this, a drive’s controller constantly works in the background, performing “housekeeping”, to recharge each cell long before they can corrupt. This housekeeping can only occur when power and resources are available, normally when a computer is turned on. Ultimately, if left turned off for long enough (10+- years for TLC drives), drives will lose their data.
You can use the tool below to charge and erase the transistor.
Transistors are made up of the following core components:
- Control Gate – The system applies either a positive voltage or negative voltage to add or remove charge to the cell in the floating gate.
- Floating Gate – inside this gate is a cell which holds charge much like a battery would.
- Substrate – this acts as a base to and from which electrons flow.
- Tunnel and Gate oxide – these act as an insulation layer that prevent the electron charges from flowing freely between the Floating gate and substrate when the Control Gate is NOT active.
- Source and Drain – these sit on either side of the substrate and are what allow the electrons to flow in and out of.
Eliza
A conversational machine from 1966
In 1966 Joseph Weizenbaum created a program; Written in MAD-SLIP for the IBM 7094 while he was part of the MIT ProjectMAC group. One of the few AI researchers in history to acknowledge computers can only be made to ‘Appear Intelligent’, that at best an intelligent machine is a ‘powerful illusion’.
The following Eliza simulations based on one of the original program modules created to mimic the role of a Rogerian psychotherapist. It became known as ‘DOCTOR’.
Eliza – mock Rogerian psychotherapist. Original program by Joseph Weizenbaum in MAD-SLIP for “Project MAC” at MIT: Weizenbaum, Joseph “ELIZA – A Computer Program For the Study of Natural Language Communication Between Man and Machine” in: Communications of the ACM; Volume 9 , Issue 1 (January 1966): p 36-45.
This JavaScript port was originally developed by Norbert Landsteiner 2005; http://www.masserk.at