Language identification system and method for a peripheral unit

U.S. Patent Number: 5392419

Abstract: A data processing system is responsive to a plurality of input languages, each language adhering to a prescribed syntax. The presence of defined portions ("For" keys) in the incoming data indicate a vote for the presence of a language and the presence of other defined portions ("Against" keys) indicate a vote Against the presence of the language. The system performs the following method for identifying the input language: analyzing, for each expected language, the syntax of an incoming block of data to identify For and Against keys in the block of data; providing For and Against tallies for each expected language in response to the analysis, each tally being a summation of key entries, each key entry comprising an identified key count multiplied by a skew, a skew value indicating the importance of the key in the syntax and in the context of said block of data, the For tally summing entries of For keys and the Against tally summing entries of Against keys; comparing the For and Against tallies to determine whether or not they are so close as to signal uncertainty; and, based upon a further syntactical characteristic of block of data, resolving the uncertainty and indicating a value based on one of the tallies: indicating a value derived from the larger of the tallies in the event of no uncertainty between the tallies; and deciding, based upon the indicated value for each expected language, the identity of a received language.

Inventors: Walton; Sean E. (Boise, ID)

Assignee: Hewlett-Packard Company (Palo Alto, CA)

Application Number: 07/825,479

Issued: 1995-02-21

Expired: 2007-02-21

Classes: 358/1.13 ; 358/1.15

Field of search: 395/500,112,114

Click the image above to view patent images at uspto.gov within a frame.

Click here for the fulltext page on uspto.gov within a frame.