Abstract: The invention relates to a method and apparatus for checking the integrity of a message such as a program and/or data against an original message supposed to correspond to it.By using an algorithm (A) on at least part of the original message, at least one signature (S1, S2, . . . Sm) of this part is calculated and memorized in a memory area (10) of a portable object (1) which also has processing circuits (11) able to implement algorithm (A). In order to check the integrity of a message, at least one signature of a message part is calculated by the processing circuits (11) of the portable object by implementing the algorithm (A) and is compared to an original message signature supposed to correspond, and memorized in memory area (10).

Abstract: A process of authentication, by an outside medium, of a portable object connected to that medium via a transmission line and a system for carrying out the process. To authenticate a portable object (7) connected to a terminal linked to the outside medium by a transmission line, the outside medium (1) manages a control table (11) containing a non-secret piece of data (N.sub.7) specific to each object that can be identified, as well as a parameter (E.sub.a ) and a result (R.sub.a) entered in the control table during the preceding authentication of the object. A new parameter (E.sub.b), produced by the medium during authentication, is used to calculate a new result (R.sub.b) in the object. The old result is recalculated. If it matches the one in the table (11), the object is authenticated, and the new parameter (E.sub.b) and result (R.sub.b) are entered in the control table in place of the preceding ones. The process is well suited for use in protecting computer networks.

Abstract: A method and apparatus for transmission of digital data in which data words having M parallel bits are encoded into code words having N parallel bits, the code words being serially transmitted and being accompanied by flags such as Start Delimiter, End Delimeter of the frame or token. The code words and the flags are selected so as to balance the DC component. The time interval between two successive transitions of the serially transmitted signals is limited to four times the period of the serial transmission clock. For detection of errors in transmission, the Hamming distance between a flag and any encoded sequence of the same length extracted from any message is at least equal to 2. The invention is useful with transmitters and receivers that exploit the properties of the transmission method, and is applicable to data networks.

Abstract: Transmission errors in an interlaced Reed-Solomon cyclically coded binary message are detected and corrected. The message is coded and decoded by a system including several serially connected shift register assemblies. The first register of the first assembly receives the message data. The number of register assemblies equals the degree of a polynomial that controls the code. The number of registers in each assembly equals the interlacing depth of the code. Feedforward circuits between the register assemblies include matrices for obtaining the coefficients of the polynomial of the cyclic code. These coefficients are obtained based on the prime elements of a Galois group.

Abstract: A magnetic recording disc includes a plurality of sectors, each of which includes a plurality of data tracks and a reference zone containing information for identifying the data tracks. Some of the sectors are standby sectors containing information which would normally be contained in the sectors containing the data tracks. The reference zones of all sectors contain binary fault indicating bits which indicate whether a particular sector contains faults, and parity bits having values dependent on the location of the parity bit and the value of the fault bit at the same location as the parity bit. A data processor commands a read out head for the disc to be located at a particular position. From the command position of the data processor and the value of the fault bit, the data processor computes an expected parity bit value. The expected parity bit value is compared with the parity bit value read from the disc to provide an indication as to whether the fault bit is accurate.