Appeal No. 2000-0810
Application 08/699,412
tables one for each candidate language; and it does not disclose using the highest
value of a raw count to determine the language or genre of a language.
Claim 15 is not much different from claims 1, 8 and 16, except that the “raw
count” feature of claims 1, 8 and 16 is replaced by a “weighted count.” Our discussion
above with respect to claims 1, 8 and 16, except for that concerning using the raw
count, also applies to claim 15.
Claim 15 additionally recites:
the words in each word table are selected based on
frequency of occurrence in a candidate language so that each word table
covers an equivalent percentage of the associated candidate language.
The examiner concludes, without citation to corresponding teachings in the prior art,
that it would have been obvious to one with ordinary skill to select words with respect to
each candidate language and maintain an equal percentage of candidate words in each
table. The examiner states that the motivation is to not favor one candidate language
over another. However, the prior art cited by the examiner does not support the
conclusion. On this record, as is argued by the appellants, the motivation of not
favoring one language over another by including the same percentage of words in each
word table stems only from the appellants’ own disclosure. Kucera does not even
disclose the use of multiple tables one for each candidate language, let alone suggest
to keep the same percentage of frequently used words in each table. The same is true
with respect to the Ejiri. The appellants correctly state that the mere fact that a
9
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 Next
Last modified: November 3, 2007