Anyway, i continue to organise other materials into something resembling web pages.
The following are things that un-doubtedly you have thought of, addressed, etc, but
anyway.

The voice generation has a long way to go. But, then that's a technology step that
other than having an actor imitate Phil's voice and do a basic phonem set (probably
something based on the international phonetic alphabet) plus a set of recorded
phrases, etc that could then be splced together to produce a more intelligble set
of phrases. this would be v. important for intenational speakers for whom English
is a second or nth language.

And on the flip side, the voice recog is obviously a long way to go (why are we
*conceptual* artists SO bogged down by such "primitive" technology -- (sounds
of techies who are v. upset that we should refer to their BRAIN CHILDS THAT
ARE IN FACT STATE OF THE ART, IF YOU DRONGO's KNEW ANYTHING ABOUT
DSP (digial speech processing; not signal).... as "primitive; oh, well).

Actually considering that the VRecog s/w *is* speaker independent, it works v.
nicely -- again, with non-native or heavily accented speakers, the VRecog will
natch have probs.

The only alternative here (and i'm thinking of researchers say in literature,
history, philosphy, etc who would be "conversing" with the PKDA as part
of their research, etc), that they would have to have "practice" sessions
for voice recog.

"Back in the days" of the early Vrecog s/w for the PC, you had to enter the
command something like 3 or 4 times (eg, "dir" (so that it could do a
DIRECTORY command in dos). This set of "patterns" was then used to
"build" up the vocaulary that could be recog'd. And i'm sure you're quite
aware of how speech changes with inflection -- i am guessing here, but
it would prob seem like a totally different word/phrase.

Anyway (again most of this stuff that i'm typing is just to bring up topics
for discussion -- my level of understanding (esp in *this* century) is that
of a slightly informed (and there-fore greatly mis-informed) amateur)...

Anyway, there are clearly several issues here (you can think of this as
something of a Dallas Morning News "Arts Section" review if you wish).

1. The PKDA as an actual android. The most offten heard comment


>
>Frank Leeding wrote:
>
>
> >Anyway, sorry i had to split so abruptly Friday, but i had a bus to
>catch.
> >
> >Also, i am v. interested in contributing at least *content* if not s/w
>to
> >the project. Are there some introductory readings i could look at, as
> >well as any sort of "programmer's guide" as to how the k-base is
> >built, etc.
> >
> >-- Frank.
> >