On 14/12/13 12:51, Steve Newcomb wrote:
> Oops. I haven't been talking about machine-machine interaction at all.
> So your point is well-taken, Dave.
Perfect example of what you were talking about. We needed more detail of
the context to understand the scope of the discussion?
...metadata?
>
> While machine/machine interaction is very interesting and important, I
> personally don't feel called in that direction, really. I've been
> interested in corporate memory, though -- the kind of memory that goes
> away when a human being dies or retires. Memory that may not even be
> logical, that may be ambiguous, and that may help when the organization
> (of whatever size or purpose) is confronted with challenges both old and
> new. In such cases, the original knower is no longer available, so
> there's no possibility of direct interaction. One-way communication
> ("one-way XML", as you say) is all there can be.
the examples being a company re-inventing X since Joe Blogs has left
who designed the original? A nominal variant on NIH?
>
> For about 5 centuries now, book publishing has been largely one-way,
> too. We've come a remarkably long way since that phenomenon began, and
> "science" -- literally, "knowing" -- has been especially favored by it.
> It's a promising avenue of endeavor, I think. For example, when I
> think about the 250 years since Darwin's astonishing book, and how the
> whole of humanity is still struggling to adapt to that particular piece
> of knowing, I am awed. And Darwin himself hasn't been around to answer
> questions for a very long time now. It's been a matter of one-way
> communication, almost all of it. Wow!
And moving even more in that direction? A guess, but many on this list
will communicate more via electronics than face to face? Which reduces
if not removes the implicit feedback of a face to face discussion.
>
> I admit that I have a perhaps-religious perspective. I just don't
> believe that machines "know" anything, in the classic sense of, uh,
> intuiting them. I see machines as the contrivances of creatures who
> actually *do* know things -- or at least who *think* they do -- and who
> use such contrivances as tools for expanding the scope of their knowings
> (or, if you prefer, their thinkings that they know).
Take the best AI machine. At best it has some subset of the knowledge
captured from the 'teacher' via the k-capture system, so I guess I'd
agree with you. Machines are dumb, given this context for 'dumb'.
regards
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk
