The LZSS compression algorithm is one of a number of compression algorithms generally referred to as "Dictionary Methods". These algorithms rely upon the fact that (in general) an input data buffer will contain repeating "patterns" or matching sequences of bytes. The algorithms fall into 2 groups. Systems like LZ78 and LZW scan an input buffer and construct a "dictionary" of the most commonly occurring byte sequences or "phrases". This dictionary is pre-pended with the compressed data and the compressed data comprises an array of indices into the dictionary. A second set is a modification of this in that the data dictionary is implicit in the uncompressed data buffer. All are based upon an algorithm developed and published in 1977 by Abraham Lempel and Jakob Ziv LZ77. A refinement of this algorithm, which is the basis for practically all the later methods in this group, is the LZSS algorithm developed in 1982 by Storer and Szymanski. These methods try to find if the character sequence currently being compressed has already occurred earlier in the input data and then, instead of repeating it, output only a pointer to the earlier occurrence. This is illustrated in the following diagram:

The algorithm searches the window (a buffer moving back from the current position in the input data). It searches for the longest match with the beginning of the look-ahead buffer (a buffer moving forward from the current position in the input data) and outputs a pointer to that match. This pointer indicates a position and length of that data match. It is referred to here as a "Slice Descriptor". Since it is possible that not even a one-character match can be found, the output cannot contain just pointers. Accordingly at times it is necessary to write literal octets into the output buffer. A block of literal octets is preceded by a "Literal Block Identifier" which indicates the length of the literal octet sequence that follows.

The following is provided as an informative example using the input buffer shown below.