Browse Prior Art Database

A (2,7) Rll Code With an Effective Lookahead of 3

IP.com Disclosure Number: IPCOM000099327D
Original Publication Date: 1990-Jan-01
Included in the Prior Art Database: 2005-Mar-14
Document File: 3 page(s) / 112K

Publishing Venue

IBM

Related People

Franaszek, PA: AUTHOR [+2]

Abstract

Described is a (2,7) fixed-length run-length-limited code whose encoder effectively depends on three contiguous lookahead symbols in the input stream. Prior art indicates that a minimum lookahead is four symbols, which the code apparently violates. But, in reality, the code is consistent with the minimum lookahead as four symbols in that it depends on the three symbols in the input stream that immediately follow the next symbol. The fact the code depends on only three contiguous symbols, not four, may simplify the implementation of the encoder, the decoder or both.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

A (2,7) Rll Code With an Effective Lookahead of 3

       Described is a (2,7) fixed-length run-length-limited code
whose encoder effectively depends on three contiguous lookahead
symbols in the input stream.  Prior art indicates that a minimum
lookahead is four symbols, which the code apparently violates.  But,
in reality, the code is consistent with the minimum lookahead as four
symbols in that it depends on the three symbols in the input stream
that immediately follow the next symbol.  The fact the code depends
on only three contiguous symbols, not four, may simplify the
implementation of the encoder, the decoder or both.

      General methods for constructing codes for run-length limited
channels are described in [1].  Such channels typically constrain
codewords so that runs of consecutive 0s have a minimum and maximum
length.  A (2,7) code, for example, is a code for which between any
two successive 1s there must be at least two 0s and no more than
seven 0s.

      Franaszek [1982] discusses a class of codes for this channel
for which encoding depends on a finite past history of states visited
by the encoder, and on the next few input symbols.  One code in this
class, the stationary code, has the property that the coding rules
are the same for each input symbol.  It is not difficult to show that
a fixed-length stationary (2,7) code must depend on the present state
and at least the next four input symbols.  The encoder may possibly
depend on the history of states as well.  The quality of a code is
measured by the length of the lookahead sequence and by the length of
the past history.  The shorter the length of the lookahead sequence,
the less costly the encoder and decoder and the more easily that the
encoder/ decoder system recovers from errors.  An incorrect bit in
the input sequence potentially causes an erroneous coding decision in
each position in the lookahead string.  Short lookahead strings
reduce the number of erroneous decisions.  Similarly, the decoder
produces fewer errors in its output strings when the lookahead string
is short.  It is also desirable for the same reasons to minimize the
state history on which the encoding algorithm is based.

      Described is a (2,7) code that seems to require fewer lookahead
symbols than the lower bound of four.  It depends on the present
state, and on three consecutive symbols in the input string.  It does
not depend on state history prior to the present state.  The reason
that it appears to break the lowe...