Compact spelling-check dictionary

• US 4,782,464 A
• Filed: 12/26/1985
• Issued: 11/01/1988
• Est. Priority Date: 12/26/1985
• Status: Expired due to Term

• Current Assignee: Smith Corona Corporation (Pubco Corporation)
• Original Assignee: Smith Corona Corporation (Pubco Corporation)
• Inventors: Duncan, Howard C. IV, Adams, Donald T., Gray, R. William
• Primary Examiner(s): Shaw, Gareth D.
• Assistant Examiner(s): Mills, John G.
• Application Number:
  US06/813,342
• Time in Patent Office: 1,041 Days
• Field of Search: 364/200 MS File, 364/900 MS File, 364/718, 273/237, 434/169, 434/335
• US Class Current: 715/259
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation

Abstract

A low-cost spelling-check dictionary apparatus for incorportation in a portable electronic typewriter. Semiconductor ROM is used for storage of the dictionary'"'"'s word listing and for program control of the spelling check. Text compression methods are utilized to permit storage of large vocabularies (about 35,000 words) while minimizing the ROM capacity required (about 3 IC'"'"'s of 256K bits each). In particular, the words are alphabetically arranged in a set of three interconnected tables together with extensions of the third table in highly compressed fashion. The tables verify the spelling order of the first three character inputs, while the extensions verify the fourth and subsequent inputs, if any. The tables have a first "look-up structure" format, whereas the extensions are arranged in a different "tree structure" type of format utilizing a "Huffman code" for the characters.

Claims

1. An electronic dictionary storage apparatus for use in an electronic typewriter or handheld dictionary, the apparatus containing an electronic processor and a memory storing a list of correctly spelled words for verifying the spelling order of character inputs comprising:

(a) input means for receiving characters from a keyboard;

(b) a first storage means for storing character information structured to sequentially check the spelling order of a first predetermined series of characters received from said input means, wherein said character information comprises a plurality of character sequences, each said sequence having a predetermined series N of characters therein and being combinable with a first character to form a sequence as a preliminary portion of one word in said list and wherein said characters are stored in ordered sets consisting of groups so that each group of set is located by a respective address associated with a particular one of said first characters;

(c) a second storage means extending from said first storage means and containing further character information structured to check the spelling order of a variable string of character inputs stemming from said predetermined series to validate spelling of words, wherein character information in said second storage comprises a tree structure with a plurality of branches, each branch having a unique address in said second storage means and each branch holding a least one word remnant associated with said character sequence;

(d) processing comparison means connected to said input means and to said first and second storage means, and being operative in response to receipt of a character from said keyboard to ascertain occurrence of a match upon comparing the received character and a corresponding stored character in proper spelling order according to said character information stored in the first and second storage means.

2. The dictionary stoage apparatus of claim 1, wherein character information in said first storage means is in a first code format and held in a plurality of bytes, one character per byte, each character associated with a respective address;

and comprises a plurality of character sequences, each said sequence having a predetermined series n of characters therein, a first of said n characters selected from all characters of the alphabet and further alphabetic characters up to an n-th character, said further characters being combinable with said first character to form a said sequence present as at least a preliminary portion of one word in said list;

said further characters are stored in ordered sets, each set consisting of at least one group of ordinally-identical higher order characters, said higher order characters in at least some of said groups being less in number than the entire alphabet;

each said group of a first said set is located by a said respective address associated with a particular one said first character, and groups in higher ordered ones of said sets are located, in turn, by the respective addresses of said further alphabetic characters selected in response to said occurrence of a match upon comparing a said further alphabetic character with a said received character; and

wherein character information in said second storage means comprises a tree structure with a plurality of branches, each branch having a unique address in said second storage means and stemming from a discrete one of said sequences by virtue of said n-th character of said discrete one sequence having said unique address associted therewith as said respective address, and each branch holding at least one word remnant associated with said discrete one sequence, said word remnant being held in a second, byte-independent code format; and

further including means converting codes in said second code format to codes in said first code format prior to ascertaining occurrence of said match.

3. The dictonary storage apparatus of claim 2, further including word terminators indicating completion of a word in said character information stored therein, and wherein at least one said sequence includes a said word terminator in a code in said first format, said second format is a compressed code format having variable numbers of bits for the different characters, some characters with bit-length in excess of a byte, said word remnant is followed by a said word terminator, and the word terminator following said word remnant has a discrete code in said second format relative to the word terminator in said first format.

4. The dictionary storage apparatus of claim 3, wherein said first code format requires less than all bits of said byte to distinguish said characters, at least one bit being in excess, and said means converting codes in said second format is normally disabled, but enabled in response to a predetermined binary state of said excess one bit of the byte storing said n-th character.

5. The dictionary storage apparatus of claim 1, further including a starting address for said word list and word terminator codes marking the end of each word in said character information stored therein, means indicating detection of a word terminator code in said character information, and an end-of-word code among said character inputs;

and wherein character information in said first storage means is held in groups by alphabetic order of association in particular sets of character combinations having a predetermined number of characters, whereas character information in said second storage means comprises a tree structure with a plurality of branches, each branch stemming from a discrete one of said sets and holding a plurality of character-containing word remnants associated with said discrete one set, each said remnant followed by a word terminator code, and still further including signal emitting means responsive to input of said end-of-word code in ordinal correspondence with a said word terminator code, and means to return to said starting address for comparing characters of a next word, said means to return being operable in response to a signal from said emtting means.

6. The dictionary storage apparatus of claim 5, further including an end-of-table code together with means detecting said end-of-table code and means responsive to detection thereof to emit a signal indicating impossibility of a match, and wherein said characters of some of said word remnants occur as separate, random-length portions hierarchically interspersed in said branch in accordance with said alphabetic order and said word terminator code following a last word remnant in each branch is followed in turn by said end-of-table code.

7. The dictionary storage apparatus of claim 1, wherein said predetermined series consists of n received characters, said character information in said first storage means comprises a succession of n ordinally-arranged look-up tables, each table containing a plurality of addresses, a particular alphabetic character present at a corresponding order in the spelling of at least one word in said list being associated with a given address of said plurality, and each address in the n-th order table pointing to location of still higher order characters stored in said second storage means, said still higher order characters forming a remnant portion of a said at least one word;

and further including access means operative to select a given one of said associated addresses in the n-th order table in response to ascertanment by said comparison means of said occurence of a match between an n-th received character and said particular alphabetic character present at said corresponding order, said given one address locating the still higher order characters in said remnant portion for comparing with further ones of said received characters.

8. The dictionary storage apparatus of claim 1, wherein said predetermined series consists of n received characters, said character information in said first storage means is in a first code format and held in a plurality of bytes, one code per byte, and comprises a succession of ordinally arranged look-up tables equal in number to said predetermined series, each said table storing a plurality of address groups, each said address group containing addresses related to a particular prior alphabetic character present in the spelling of a word in said list and each said address in a given address group pointing to the location of a group of next higher order characters stored in a next one of said tables;

each said next higher order character is associated with a respective further address and combinable with each lower order character associated with a prior said address pointing to the location of the group containing said next higher order character, character combinations form a plurality of n-character sequences, each sequence present in the initial characters of at least one properly-spelled word of the list, and the ordinally-identical characters in at least some of said groups are less in number than the entire alphabet; and

wherein said n-th table stores groups of n-th order alphabetic characters, each n-th order alphabetic character of a said group being associated with a respective output address, said output address locating in said second storage means the start of character information including a given one of said n-character sequences in forming words of the list;

character information in the second storage means is in a second, byte-independent code format and comprises a tree structure with a plurality of branches, each branch stemming from a discrete one said sequence and holding at least one word remnant associated with said discrete one sequence, and further including means converting codes in said second format to codes in said first format prior to ascertaining occurrence of said match.

9. The dictionary storage apparatus of claim 7, wherein n is equal to three, said first storage means comprises three look-up tables for defining a path to said second storage means, a first one of said three tables containing a stack of addresses, each address relating to a character of the alphabet as a first character of a sequence spelling at least one word of said list and pointing to a next one of said tables storing secondary addresses in discrete groups;

each said secondary address in a said group is associated with a respective second alphabetic character combinable with said first character to form a unique bi-character sequence present in the spelling of different words of said list having said first character in common;

said second alphabetic characters serve as arguments in said next one table to select a respective one of said secondary addresses, each said secondary address pointing to the third said table, said third tablle storing tertiary addresses in discrete further groups, each said tertiary address in a said further group likewise associated with respective third alphabetic characters combinable with said first and second characters to form a unique tri-character sequence spelling a preliminary portion of particular ones of said different words of the list;

said third alphabetic characters serve as arguments in said third table to select a respective tertiary address locating pertinent character information in said second storage means, said pertinent information containing a plurality of characters, at least two characters of said plurality each completing a distinct word of said particular ones of the different words of the list and sharing said unique tri-character sequence as precedent characters; and

wherein selections of said respective secondary and tertiary addresses occur in response to said occurrence of a match between a said corresponding character and a second and third of said received characters, in turn, whereby said pertinent information in said second storage means stems from said unique tri-character sequence in said first storage means in response to selection of said respective tertiary address.

10. The dictionary storage apparatus of claim 7, wherein numerals are associated with a plurality of said word terminator codes in said second storage means and signify in said word list the nodal points where next higher-order characters of a like plurality of others of said remnants each follow a said numeral and share precedingly matched characters, and further including detecting means for said numerals, a register, and a counter defining the ordinal position of each received character and initially at a unit value, said counter being incremented to a new value in response to receipt of each said character input;

control means operative in response to non-occurrence of a said match between a received character and said corresponding stored character, said control means storing the new value in said register and initiating a scan of subsequent character information for said numerals;

means supplying said comparison means with the new value and a detected numeral in response to detection of a numeral by said detecting means, means responsive to identity between the new value and the detected numeral upon comparing same, said last-mentioned means being operable to compare said received character and said next higher order character following said detected numeral, and means repeating said scan in response to further non-occurrence of said match upon comparing the received and next higher order characters.

11. The dictionary storage apparatus of claim 10, further including a subtractor having a borrow indicator and wherein said subtractor performs a subtraction of said new value stored in said register from each said numeral detected by said detecting means, said subtraction occurring prior to supplying said comparison means with the new value and said detected numeral, and further including control means for halting said scan in response to said borrow indicator.

12. The dictionary storage apparatus of claim 7, further including word terminators indicating completion of a word in said character information stored therein, and wherein at least one said sequence includes a said word terminator as a code in said first format, said second format is a compressed code format having variable numbers of bits for the different characters, some characters with bit-length in excess of a byte;

said word remnant is followed by a word terminator, and said word terminator following said word remnant has a discrete code in said second format relative to said word terminator code in said first format.

13. The dictionary storage apparatus of claim 12, wherein said first code format requires less than all bits of said byte to distinguish said characters, at least one bit being in excess, and said means converting codes in said second format is normally disabled, but enabled in response to a predetermined binary state of said excess one bit of the byte storing said n-th character.

14. The dictionary storage apparatus of claim 7, further including word terminator codes marking the end of each word in said character information stored therein, means indicating detection of a word terminator code in said character information, and an end-of-word code among said character inputs;

and wherein character information in said first storage means is held in groups by alphabetic order of association in particular sets of n-character combinations, and character information in said second storage means comprises a tree structure with a plurality of branches, each branch stemming from a discrete one of said sets and holding a plurality of character-containing word remnants associated with said discrete one set, each said word remnant followed by a word terminator code, and still further including signal emitting means responsive to input of said end-of-word code in ordinal correspondence with a said word terminator code, and means to cease said comparing in response to a signal from said emitting means.

15. The dictionary storage apparatus of claim 14, wherein an end-of-table code is present in said branch, said characters of some of said word remnants occur as separate, random-length portions hierarchically interspersed in said branch in accordance with said alphabetic order and said word terminator code following a last word remnant in each branch is followed in turn by said end-of-table code;

and further including means detecting said end-of-table code and means responsive to detection thereof to emit a signal indicating impossibility of occurrence of a match.