Appeal No. 2000-0810
Application 08/699,412
language, having a list or table in the first instance for a second language is not a mere
duplication of what existed before. The examiner has not articulated and established a
motivation for one with ordinary skill in the art to expand Kucera’s system to support
multiple languages and to support them by having a separate word table of most
frequently used words for each.
Third, each of claims 1, 8 and 16 calls for accumulating a respective count for
each candidate language when there is a match with a word in the corresponding word
table, and identifying the language as that associated with the count having the highest
value. The appellants argue that these claims require the use of a “raw count” or “non-
normalized sum.” The examiner disputes that appellants’ claims specify identification of
the language used on the basis of a raw count or non-normalized value and thus
ignores this limitation. But at least with respect to claims 1, 8 and 16 and claims which
depend from claims 1, 8 and 16, the highest value of the accumulated counts for each
table determines and identifies the language. That means raw count or non-normalized
value is indeed used to make the language identification. The examiner’s view is
without merit and has provided no sufficient basis to refute the raw count or non-
normalized value argument of the appellants.
Ejiri is relied on by the examiner for its statement that the frequency of a word or
n-gram has been used as a clue to identify an author or language. That general
disclosure does not remedy the deficiencies of Kucera regarding the appellants’
claimed invention. It does not disclose the use of multiple tables, much less multiple
8
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 Next
Last modified: November 3, 2007