A machine language for documentation and information retrieval

In transcribing documents for machine entry, the font limitation of existing input devices becomes a barrier to full information flow. The printer's press has more than 3,000 symbols commonly available; a Flexowriter has less than 90, a card punch has less than 50. A satisfactory input language must significantly reduce this disparity and meet three additional requirements:1) It should require a minimum of intellection on the part of the input typist.2) The typed copy should be readable or readily translatable.3) The language must be machine interpretable by unequivocal and simple algorithmsA method is shown meeting these requirements, which uses pseudo-alphabets having mnemonic names, wherein the initials of these names are used in a manner analogous to single-address computer codes. The sequence of a transform signal followed by an alphabet initial specifies the operation or transformation, and simple format rules specify the extent of the operand. The most frequently used elements of these pseudo-alphabets have themselves a high degree of mnemonic correspondence to the keyboard symbols that represent them. The method is shown to be machinable and the logic of a simple interpreter-translator is presented.