MW: Images are *not* an efficient way to store data. Unless they are
three-dimensional images, they lack data. Normally, they include a lot of
unnecessary or redundant data. It is very, very rare that a computer
stores any but the smallest image without compressing it. And remember,
an image can be stored as symbols in a relational database very easily as
a set of x-coords, y-coords, and colors.
MD: So why store the data in an image in the first place?
GENERAL POINT: Mike D takes MW's argument to its logical extension. Why have
images at all? If symbols are so efficient... Why not just have "pi r 2" [pi
r squared] and leave the circle out of it? And similarly for every other
shape. Why have ANY images? Let's just have words and numbers. One or two
symbolic languages, and that's it. On the face of it, it would be so much
simpler. And philosophers have been dying & trying to achieve just one
language for thousands of years.
And yet that's not possible, is it? (Or by all means, be my guest & try to
argue and demonstrate that it is).
But you guys can't explain why we DO need images and graphics, can you? ...
Can you? And that should give you pause.
The reason we can't make do with symbols alone is that they are abstract.
They bear no resemblance to their objects at all. Only you literally can't
"see" that. You think that symbols do magically depict their objects. For
you "pi r 2" IS a graphic of a circle. Because your brain automatically
supplies that graphic..
The simple way to demonstrate that symbols don't depict things is to change
the language. What are "pce" , "rtf (3)" "gh-jc / xt"? These are three
notations I've just invented for a square, hexagon, and circle. But which is
which? There's no way you can tell - unless you have the "sense code" - i.e
know what images or graphics they refer to .; otherwise you're lost. And you
would be lost with existing mathematical notation, if you hadn't already
learned the sense code.
Same thing with ordinary language. If the program is written in Bulgarian,
or some language you don't know, it's not going to mean anything to you. If
it tells you to "rute" then "grote" then "itumise".. it's a case of wtf?
But there's no such problem with images and graphics - they are a universal
"language" - no one of any nationality has any problem understanding a
photograph (unless it's taken from a weird angle - which they've never
*seen* before) or a triangle or circle. Because images & graphics are more
or less concrete - isomorphic with their objects.
We can't make do with just symbols alone - we'd be utterly blind.
And the part that you're overlooking and missing out in all your
arguments.about how computers handle images, is that in the end, after first
re-expressing them as symbols, those symbols then always are and have to be,
converted BACK into images and graphics. (You have to .e.g., see the route
you're supposed to take, you can't just work with a math formula - or, pace
MW, a complex set of x-,y- coordinates).
As Richard L v. succinctly put it in a post recently on Semantics, computer
programmers over and over, analyse computer operations simply in terms of
what they overtly do, and overlook that these operations only work because
the programmer's brain is implicitly providing additional information, that
in this case makes sense of them.
So images and graphics are a FUNDAMENTAL part of all the operations we're
discussing - just not always obvious.
EFFICIENCY OF STORAGE: MW made a lot of points which I found interesting
about how computers handle images, but he didn't answer the first question :
isn't a graphic, like a map, particularly of a complex form like, say Italy,
a simpler form of representation, than its expression in a relational
database? From Charles Hixson's answer, I take it - but correct me - that it
is. (I'm pretty sure it doesn't need repeating, but obviously I'm a
techno-idiot, v. ignorant about these computational matters).
OPERATIONS THAT CAN BE CONDUCTED ON GRAPHICS: The reason I keep reworking
certain arguments is that while the need for graphics is intuitively obvious
to me, it's v. difficult to analyse and especially put into words. And I'm
still groping here.
But manifestly there are many operations on physical forms that computers
can't handle, when they are obliged to work only on the symbolic expression
of those forms.
Yes, for example, computers can morph a face to aid a plastic surgeon, and
they can I would imagine, suggest standard ways of altering eyes and other
features - that such and such features might be suitable for a "George
Clooney" or a "John Travolta" look. But they can't create AND judge a
different look - can't freely mix a bit of Clooney and a bit of Travolta and
decide that yes, that will work, or no it will be a disaster, as humans do.
For that kind of aesthetic judgment, I suggest, you have to work directly
with shapes.
They can't IOW be designers. They can only alter shapes in narrow AI ways,
along set, formulaic lines. Designers can and do take given letters, for
example, and produce a practically endless set of "out of the box"
variations on them - think of the vast variety of logos humans produce,
which radically alter the components of a given shape, like a letter, and
yet still preserve a likeness. A designer, for example, convert a
straight-line A into a set of circles, which are still recognizable as an A.
That can only be done, I suggest, by working directly with shapes/ graphics.
There is no formula for inventing a radically new-shape "A" or, for that
matter, chair. The essence of creativity is that you break the rules.
(We're talking here, BTW, about the difference between narrow visual AI, and
visual AGI).
Secondly, they can't, I suggest, draw analogies. They can't as the human
mind, produce a series of analogies like "rope", "snake," "chain," and
"spaghetti strand". Or think of the sky as "crying" - drawing an analogy
between a raindrop and a tear. That, it seems to me, is obviously done by
comparing shapes directly, not by comparing their symbolic expressions.
MW considers this a crackpot idea. Well, can we first agree on something
here? - the human mind can indeed produce such a series of analogies, and
do so with very few operations, very quickly. We can't obviously be sure how
much is going on unconsciously in the brain there, but it seems likely that
the brain can produce such "fitting" comparisons, with only say a hundred
operations or so. Fair estimate? The brain clearly doesn't search
systematically and extensively through a database the way computers do.
So how can this be done - with so few comparisons - by working on any kind
of symbolic database? The brain is never in danger of combinatorial
explosions in its searches, is it?
.
-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=8660244&id_secret=56004435-501149
