Browse Prior Art Database

Error Correction as an Extension of Error Recovery on Information Strings

IP.com Disclosure Number: IPCOM000052605D
Original Publication Date: 1981-Jun-01
Included in the Prior Art Database: 2005-Feb-11
Document File: 2 page(s) / 58K

Publishing Venue

IBM

Related People

Gourneau, DA: AUTHOR [+2]

Abstract

When errors occur on a read operation such as reading diskette files, retries are attempted until the record is recovered or the retry count is reached. If an error exists after the retry count has been reached, the error is considered hard and the record is unrecoverable. As illustrated in the drawing, a read verify operation is performed after the writing of data on diskette to validate the integrity of the data record and the cyclic redundancy check (CRC) bytes.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 74% of the total text.

Page 1 of 2

Error Correction as an Extension of Error Recovery on Information Strings

When errors occur on a read operation such as reading diskette files, retries are attempted until the record is recovered or the retry count is reached. If an error exists after the retry count has been reached, the error is considered hard and the record is unrecoverable. As illustrated in the drawing, a read verify operation is performed after the writing of data on diskette to validate the integrity of the data record and the cyclic redundancy check (CRC) bytes.

The data which has been written is usually read at some later time during a normal read operation. The data read is checked and if there is a data check, conventional retries are attempted. If during a retry a successful read occurs, the record is considered good. If all the retry attempts are unsuccessful, the error correction extension operation is effected.

The error correction extension operation includes the reading of the data record and CRC bytes and calculating a remainder. Successful read operations result in a zero remainder, but since a data check had occurred, the remainder probably will not be zero. The possibility of a zero remainder does exist, and hence a check is made. A further requirement of the present error recovery procedure is that the remainder be the same on three consecutive reads. This is done because the characteristics of some diskette CRC errors are such that the number of bits in error is not consis...