The word "computer" has evolved over time to refer to a device that uses electronics to achieve computation. Prior to the second world war, most computing devices were analog, but also the word "computer" was defined as a human who performed calculation, often with the aid of machinery. The Computer History Museum in Mountain View, California has an excellent summary of the era of human computing. A classic example of human computing is the production of books of mathematical tables containing values of functions such as sine and cosine. Less obvious, but no less intriguing, are computations that occur regularly whenever we get together and collaborate. For example, two people working with each other in the kitchen to make dinner is a computation resulting in a meal. The collection of human collaborators and their material flows is a type of automaton.
I was in New York City last week and stopped by the new Museum of Math (MoMath at 11 East 26th Street in Manhattan -- not too far from the Empire State Building). A worth while visit, and many excellent exhibits. One exhibit is called "String Product," pictured above from the New York Times Alice in Math Wonderland article announcing the museum's grand opening back in December 2012. String Product is an analog computer that calculates products, such as a times b, where a is on one circle of the sculpture, and b is on another circle with a string connecting a to b. At the center of the structure, from bottom to top, numbers are listed in order, and the ab product can be found by noting where the string crosses the center. The strings are chords of the structure, known as a paraboloid. A subsequent Times page provides additional information on the museum, and includes a description of how the string product functions.
What is the connection between automata and data? Think of data as the material that flows through the machine. The data may be unstructured or structured but in either case, data can be "serialized." Popular approaches to serialization are XML and JSON. In XML, for instance, you can create a stream of characters that contains arbitrary structure (tree, graph). This stream enters the automaton, is processed by it, and leaves the machine changed or transformed in some way. Sometimes, we think of "information" as being equivalent to "data," but this is not so since information, as with the serial encoding of XML, can encode any structure from data and formula to model and program. In the old days of writing code, we used to write a computer program and the data it accessed with hexadecimal characters (0 through F): programming in bits. A cursory glance would not indicate which bits were data and which were code. The only computer I could afford after getting a bachelors degree was the Cosmac VIP sold by RCA, and I programmed using code similar to the image in a language called CHIP-8. We are now in the age of Big Data, but without automata to process that data, we are left without code, models, machines, automata, and a way forward.
What has two personalities and sits on your desk? A computer. To do in your spare time: search for the word "automata" in an image-based search. You'll probably see some crafty wooden machines and then a few odd looking diagrams with circles and letters. This is the split-personality surfacing. Prior to 1950 the word automaton, and the word computer, either meant a human being (who computed) or an analog machine. This is not strictly true because, for example, Babbage's 19th century engine designs were digital; however, up to World War II, analog computing was in wide use making digital computers relative newcomers. In the above image, on the left we have a deterministic finite state machine represented as a diagram. On the right, one of Cabaret Mechanical Theatre's inventions (Pirate Panic). Pirate Panic is driven by an analog computer, and like most analog computers, it solves a formula resulting in a continuum of states. Thankfully, the formula is most entertainingly realized with a Pirate and an Octopus. So, will the real automaton come forward? Both are valid automata and the divergence of types could be covered adequately only in a book. The short story is that the mathematical theory of automata proceeded in one direction, and the analog variety lives on in disciplines that rely on "signals and systems" which interestingly enough, means some computer music interfaces and languages in addition to the Pirate.
Machines can take on many forms, including this one which is flat and would be formally described by a network of activity nodes and arcs among them. The above example is a machine (or automaton) that is defined by flow, as are all automata. The automaton models the censorship process in Iran beginning with the Supreme Leader and then cascading down to the enforcers. This machine is constructed by The Annenberg School for Communication at the University of Pennsylvania. There is a complete description of the process, with the above figure at its original resolution at their web site. Nodes are points of control and arcs probably carry both control and data flow. Curiously, there is very little feedback in this architecture (only at the very end). There is also an interesting correlation between the abstract semantics of machines and Iran's censorship process: power flows from top to bottom (both literally and figuratively). Machines of this sort are a bit different than those in some of the other posts because they are often called models--specifically simulation models. To be more precise, this is a simulation model of the censorship process.
I had a discussion online with Ray Winstead who recently retired from the Indiana University of Pennsylvania in Biology. Ray has a page which I came across and found a diagram which is most compelling both for its biological meaning as well as the design (created by William Standaert):
The design is a flow model which receives energy in the form of sunlight and this energy undergoes a "food chain" transformation. The transformation begins with producers (e.g., plants) which are eaten by herbivores, which are in turn eaten by carnivores, to end the chain with the top carnivores (e.g., humans). There are actually 2 sub-chains in this diagram: one for consumers of live material, and the other for detritus. These types of models can be simulated: common model types used by ecologists include Odum graphs, System Dynamics, and compartmental models. Further information with a more abstract model can be found in the Wiki link on energy flow.
This is the creative automata blog for the Creative Automata Laboratory at the University of Texas at Dallas. The goal of the lab is to explore representation of process abstractions used in mathematics and computing starting with historical automata up to theoretical automata, code, and data in present day technology. Representation is informed by areas outside of computer science such as design, the arts, and humanities. The purpose of the representation is enhanced mass communications, education, and training in the arts of computing. Paul Fishwick serves as Director of the Laboratory.