Question #1

 

Succinctly describe the MEMEX. What was its purpose?

The MEMEX was proposed by American scientist Vannevar Bush in 1945 for the purpose of solving the growing problem of information overload. As Bush put it, "we can enormously extend the record; yet even in its present bulk we can hardly consult it." Bush designed the MEMEX to be "a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility. It is an enlarged intimate supplement to his memory."

In its physical aspect, Bush envisioned the MEMEX to be "a desk...the piece of furniture at which [the researcher] works." The desk would be equipped with translucent screens for viewing, a keyboard and transparent platen for entering input, and buttons and levers for manipulating the data storage medium, microfilm, which would be produced, stored, edited, and processed in the bowels of the desk.

The microfilmed information that was to be the library of Bush's MEMEX could either be bought/sent as a gift (as books are) or self-produced. Each frame/set of microfilm would have a unique identification number that could be stored in a code book for instant retrieval.

In addition to addressing the problem of information overload, Bush also wanted the MEMEX to mechanize the repetitive elements of creative thought, just as adding machines were mechanizing numerical computations and computers were mechanizing logical operations. The MEMEX would thus serve as a time-saving device for intellectuals.

 

What is the essential feature of the MEMEX?

In Bush's words: "The process of tying two items together." In other words, linking.

Bush wanted a machine that would work as closely as possible like the way the human mind worked. The human mind, of course, does not work by making a hierarchical index of everything it knows; it works by making associative links between bits of information (one could argue, in fact, that creativity can be defined as the ability to make such links).

The MEMEX would work by the researcher building associative trails through the data he had stored in his microfilm library. Each trail would be a uniquely identified series of links which denoted connections between the two records linked. By following the trail, a researcher could follow a single idea across the breadth of human knowledge, as if he were an adventurer on an intellectual safari.

Bush foresaw that these trails -- these series of links -- could be exchanged between people as easily as other data. A researcher could photocopy (literally, in Bush's system!) the trail and the documents in it and send the entire bundle to another researcher. Trails could be bought, ready-made, as we buy encyclopedias (one is reminded of the BBC show "Connections").

And trails could be saved as easily as data could. If the scholar wanted to review an old line of reasoning, he merely needed to call up his trail from a code book. And after the scholar's death, his reasoning would survive as well as his results (in Bush's metaphor, the scaffolding as well as the structure).

 

What is the hypertext model of the MEMEX?

Bush's model of hypertext had four elements: microfilm records, unique identifiers, links, and associative trails.

Each microfilm record (or related set of records) was given a unique identifier, which was imprinted in photocell-readable dots at the bottom of the record. This identifier was stored in a code book, and the researcher could call the associated record up by entering the identifier on the keyboard.

A researcher could link two records together by bringing the records up on two adjacent screens, then tapping a button to fill a code space at the bottom of the first record with the unique identifier of the second record. From that point forward, the researcher could access the second record by clicking the proper button while viewing the first record.

The links thus formed were uni-directional, and could be themselves linked together into associative trails. These trails themselves could become data, be assigned a unique identifier, and be stored in a code book.

 

Go to next question