|
Authored by: Anonymous on Friday, July 20 2012 @ 12:13 PM EDT |
Not in the sense of what's "required" for a software patent to be
granted!
Simply expressing a computer generated voice - without actually
getting into the details - is sufficient.
Caveat: I absolutely do not
agree with that decision nor the fact that the USPTO grants patents based on
such. I simply point out that with how patents are granted today, identifying
voltage and such are no longer required.
RAS[ Reply to This | Parent | # ]
|
|
Authored by: The Cornishman on Friday, July 20 2012 @ 03:37 PM EDT |
Algorithm disclosure does not seem to be required (caveat: I am taking PolR's
evaluation of Fonar at its face value, I haven't read the case).
It seems to me to be sufficient to say "a computer program samples
microphone output and parses the patterns of amplitude into phonemes. These
phonemes are assembled into candidate words with specific probabilities;
semantic contexts of the candidate words are used to distinguish between
probable meanings. Utterances are compared to databases of previously
encountered utterances and where ambiguity is found the program prompts the user
for disambiguating utterances".
Are we saying that a functional description of that sort is good enough? Does
the patent have to disclose *algorithms* for phoneme identification, word
boundary separation, semantic construction, etc.? I'm asking, wanting to know!
---
(c) assigned to PJ
[ Reply to This | Parent | # ]
|
|
|
|
|