System for compressed storage of 8-bit ASCII bytes using coded strings of 4 bit nibbles

A method of storing alphanumeric text in a random access memory where the text is received as a string of characters, each character being coded as a standard 8-bit byte, comprising the steps of:

• separating the string of characters into identifiable groups of one or more characters per group, comparing each selected group of characters with a first addressable table of stored groups of characters to determine if a corresponding group of characters is present in the table, storing a first 4-bit unit in said memory coded to indicate the entire group of characters is also present as a group in the first table and storing at least one additional 4-bit unit in said memory coded to indicate the address of the group of characters in the first table in the random access memory if a comparison between the selected group of characters and a group of characters in the table is found, comparing a first beginning portion of the selected group of characters with a second addressable table of stored groups of characters if all the characters of the selected group do not compare with any group of characters in the first table, storing a 4-bit unit coded to indicate that the first beginning portion of the selected group of characters in the second table and storing at least one additional 4-bit unit to indicate the address of the group of characters in the second table in the random access memory, comparing a second beginning portion of characters in said group of characters with said second table only if said first beginning portion is found in said second table, storing a 4-bit unit coded to indicate that the second beginning portion is in the second table and storing at least one additional 4-bit unit to indicate the address of the group of characters in the second table, comparing an ending portion of the selected group of characters with a third table of stored groups of characters if all the characters in one of said beginning portions do not compare with any group of characters in the second table, storing a 4-bit unit coded to indicate that the ending portion of the selected group of characters is in the third table and storing at least one additional 4-bit unit to indicate the address of the portion of ending characters found the third table in the random access memory, storing the number of any remaining characters after any portions found in the second and third tables have been identified or not found in the tables, and storing the remaining individual characters in coded form as one or more 4-bit units in the random access memory.

A method of encoding alpha character words in which the individual characters of the words are initially coded in ASCII coded 8-bit bytes, comprising the steps of:

• storing a set of word beginning character combinations in a prefix memory, comparing the beginning characters of the alpha word being encoded with each of the beginning character combinations in the prefix memory to determine if the same combination of beginning characters of the alpha word are present in the prefix memory, generating and storing a binary coded value indicating that the beginning combination of characters of the alpha word are in the prefix memory and indicating the location in the prefix memory if the beginning characters are present in the prefix table, deleting the beginning characters from the alpha word if they are present in the prefix memory, storing a set of word ending character combinations in a suffix memory, comparing the ending characters of the word being encoded with each of the ending character combinations in the suffix memory, generating and storing a binary coded value indicating that the ending character combination of the alpha word is in the suffix memory and indicating the location in the suffix memory, deleting the ending characters from the alpha character word if they are present in the suffix memory, storing at least one set of all possible alpha characters in a predetermined sequence in a character memory, generating and storing a binary coded value indicating the number of characters remaining in the alpha word being encoded after any prefix and/or suffix characters have been deleted, and generating and storing a binary coded value for each of the remaining characters in the alpha word, the coded value indicating the location of the alpha character in the character memory.

A method of decoding a compressed word of information stored as a string of 4-bit nibbles into a set of ASC11 coded 8-bit bytes wherein the value and relative sequential position of the respective nibbles in the string is used to select the bytes, comprising the steps of:

• determining the value of the first nibble in the string;
• if the first nibble is a first predetermined value, using the second nibble to address a first table of ASC11 coded character combinations forming the beginning of words;
• storing the addressed combination of characters from the first table as the prefix of the word;
• if the first nibble is a second predetermined value, using the second nibble to address a second table of ASC11 coded character combinations forming the ending of words;
• storing the addressed combination of characters from the second table as the suffix of the word;
• if the first nibble is either said first or said second predetermined value and the third nibble is a third predetermined value, using the fourth nibble to count out a corresponding number of succeeding nibbles;
• using each of said succeeding nibbles to address a table of individual ASC11 coded characters;
• and
• storing the addressed characters from the table as the stem of the word.

Apparatus for compressing English text coded as a sequence of ASCII coded 8-bit bytes as strings of 4-bit nibbles comprising:

• means storing a group of bytes corresponding to one word of text,
• means for comparing the beginning bytes of the stored word with groups of bytes stored in a prefix table,
• means responsive to the comparing means for signaling whether or not the group of bytes is in the prefix table,
• means responsive to the signaling means for generating at least two nibbles identifying the location of the beginning group of bytes in the prefix table,
• means removing the beginning group of bytes from the stored word of text,
• means responsive to the signaling means when the beginning bytes are not in the prefix table for comparing the ending bytes of the stored word with groups of bytes stored in a suffix table,
• means for signaling whether or not the ending group of bytes is in the suffix table,
• means responsive to the signaling means for generating at least two nibbles identifying the location of the ending group of bytes in the suffix table,
• means removing the ending group of bytes from the stored word,
• means responsive to the signaling means when the ending bytes are not in the suffix tables for counting the number of remaining bytes of the stored word,
• means generating a nibble indicating the counted number of remaining bytes,
• means responsive to each of said remaining bytes in sequence for comparing each byte with all characters stored in a character table,
• means responsive to the comparing means for generating at least one nibble for each byte identifying the location of the same byte in the table of characters, and
• means storing each of said generated nibbles in predetermined sequence.