Compression and Compaction

Compression removes strings of repeated characters, thereby reducing the amount of data that must be sent over communication lines to the remote work station. Deleted characters are replaced with a one- or two-byte character string, the first of which is called a string control byte (SCB). If a string of two or more repeated blanks is removed, a special SCB with a count indicating the number of removed blanks is inserted. The count can indicate one to sixty-three. Strings of three or more repeated nonblanks are replaced by a two-byte character string. The first byte is an SCB indicating a repetition of a nonblank with a character count. The second byte is the value to be replicated. Since SCBs can look like a data byte, noncompressed character strings must also be preceded by an SCB. This allows the receiver to locate SCBs and carry out decompression SCBs. They indicate a noncompressed string, a string of blanks, or a string of nonblanks. Compression is very useful if the original data contains large amounts of repeated characters.

Compaction is similar to compression in that it is an attempt to reduce the number of characters sent over communication lines. To use compaction, you must first define one or more compaction tables. A compaction table defines the set of characters which can be included in a compacted string. Both master characters, which can be represented using a 4-bit code, and nonmaster characters, which can be represented using an 8-bit code are included. The compacted character string begins with a compaction SCB.

A single file can be both compressed and compacted. However, each character in the file is represented in either a noncompressed, compressed, or compacted character string.

Compaction tables are defined by using the COMPACT initialization statement. The SYSOUT statement can be used to specify that certain SYSOUT classes are to be compacted and which table is to be used. The user can also request compaction and specify the name of a compaction table in the JCL through the //*FORMAT statement.

Compaction can reduce the number of bytes sent out by up to 50%; however, the outbound data would have to be completely composed of master characters to reach the full 50% reduction.

If at the time the session is established, the work station requests that ASCII be used rather than EBCDIC, then PS in JES3 translates all data to/from ASCII from/to EBCDIC. If ASCII is used, compression, compaction, and certain SCS control codes are not used.