Intro -- methodologies
Mood:
caffeinated
Topic: General info
This blog is a set of conversation extract cells
(xcells for short)
The idea is that there are blocks of conersation
that are triggered (some how)
QQ: What is the moon made of?
ANSWER (mood=humorous):
Green cheese.
END-ANSWER
ANSWER:
Mostly rock. It doesn't have an atmosphere after all. Well not much of one.
LINK: What do you mean?
Well, it has an atmosphere of mostly Argon gas
(which is produced by radioactive decay). Argon
is a fairly heavy element and so the moon's
week graviational pull can hold onto that.
END-LINK
LINK: So you couldn't survive there?
No, you'd die of space vacuum. It's atmosphere
is much thinner than even that on top of Mt.
Everest.
END-LINK
END-ANSWER
And as each part is triggered (probably at
random) it then isn't re-triggered (later)
if the same question (in the same context)
arrises again. But, a TICK mark is noted
that that particular was executed. And they
ask again, you might get a response:
Weren't you listening? I said the moon is made
of green cheese [this response is constructed
based on the setting of the MOOD variable
and the response in it -- which is extracted
and contracted.
Again, this goes back to building a believable
system. And since it is deterministic (see
the comments on the Schatzmann, et al page)
then eventually it has to repeat itself or
respond "well, i don't know anything else to
say".
One idea is also, that based on context KEY words
the topic can be brought back up. For example,
if we are seriously discussing the moon. and some
how it comes up to something like "yes, but
people used to think the moon was inhabbited",
then the system could say, "like i sed, it's
made of green cheese"
then the techs could expand that bit later with
a new link
LINK: I wonder how that rumor got started?
I read somewhere that professor nimnull discovered
an ancient green-cheese text in the himalayas...
END-LINK
The idea that the dialogs are dynamic and can
grow and that eventually there are several diff
responses in a topic area.
So, we go back to: Why do we have conversations
at all?
more later.