Content and Random Addressable Memories

Self-learning failure-proof mass memories for brain-like archiving systems

Application Proposal: Old seminal proposal from 1997

Abstract: The CAROM is a new storage device that challenges the very concept of “memory” as we know it. Traditional memories store unrelated data bits according to Shannon’s information theory. In contrast the new Autosophy information theory treats information as nodes in a mathematical hyperspace. This produces both very high lossless data compression and virtually instant search access independent of database size. Data storage is self indexing and hyperspace saturating so that the more information already stored the less additional memory required to store additional information. The new memory device consists of very thin foils with printed polysilicon thin film transistors, diodes and fuses. Information is stored by setting non-volatile fuses on the foil. The foils are wound into spindles, resulting in very rugged packages similar to paper capacitors. Other important features include: very low production cost; high memory capacity; all solid state construction; light weight; high speed random access; both random and content addressable operating modes; extreme resistance to radiation and temperature variations; low power consumption; and high reliability due to self-repair and self-healing capabilities. The new memory is suitable for both mobile applications and very large databases.

Anticipated Applications: A printed circuit module plugged into the PC bus would allow real time secure teleconferencing via the Internet. All types of data including text, fax, high resolution still images and live video are highly compressed and encrypted. The new (64k by 25 bit) Content Addressable Memory can be contained in a single chip for inexpensive mass production. A second application would supersede the CD-ROM and DVD with a solid-state credit card-sized module requiring no turntables or lasers for playback. Each module could store whole books or full-length digital movies. A self-learning no-programming Autosopher for self-organizing databases and intelligent self-learning robots would require Terabit-sized memories. Such memories would consist of spools of stainless steel foil, the size of a roll of toilet paper. Inserted into sealed capsules, the memory would be virtually indestructible and even able to self-heal and self-repair. Self-learning electronic Autosopher using the CAROM will eventually replace the programmed data processing computer.

Keywords: CAROM, DECAM mass memories.

Available downloadable documents:

Proposal document – MS Word doc

Related Publication 2005 – Webpage htm

Related Publication 1997 – Webpage htm

Patent 5 576985 -- pdf