>You could phrase it another way: blue mailboxes are "real" only to
>people in the USA.
True. Mailboxes are a bad example of icons.
>That being said, note that people the world over have pretty much
>agreed on "envelopes with stamps on them" as really being mail. It's
>a code established quite awhile ago, but it's still just a code.
>There's no reason in principle that people couldn't have decided to
>ship "mail" inside of hollowed-out gourds, and thus to say that that's
>what mail "really is." Gourds might have practical advantages for
>shipping in many climates. But instead we have rectangular envelopes
>with rectangular stamps on them, and rectangular packages with
>rectangular stamps on them, and an international mail system. So
>that's what mail "really is," for us.
>
>So I would suggest that an "icon," in your usage, is really just a
>picture of a very commonly understood code. There really aren't any
>icons, just codes that different groups of people recognize more or
>less readily.
Hm. To go further, a 'icon' is a coded reference to a reference, like
letters. Wheras 'code' is a representation of something non-physical (or at
least non-visual). Based on that I would claim that there are 'icons' (at
least view from the definition in my dictionary - 'a symbol representing or
analogous to the thing it represents') Icons _do_ point to something real,
but the refered might be a metaphor itself. (Seems like we agree somewhat,
I just a bit sloppy with the terms)
>
> The problem for cyberspace architects (and other authors) is that there is
> very little precedence to build a spatial vocabulary on.
>
>The Anthropologist's answer to that is one word: "rubbish." :-)
>
> For instance, how does one
> make a object to signify 'link' that is readily understood by most
> visitors?
>
>Give them a wall with a doorway to enter. People all over the world
>understand walls with doorways through them. Or a window in a wall to
>look through, so they can see what's beyond.
>
> You can't =8A yet; until a spatial language is formed this will be
> impossible.
>
>Human beings already have plenty of codified spatial languages. The
>trick is to use codes that are general enough, cross-culturally
>speaking, that lots of people can understand them. As long as you've
>got widely recognized codes available, you might as well use them.
I stand corrected. My posting is vague and unclear at some points. I do
know that there are a lot of spatial 'codes', readily understood, at least
within their cultural sphere. My point was that there is a lack of
cyberspace specific spatial codes, the subleties. Of course you can
represent a link in a HTML like world by making a door, no problems for
no-one. But the link is no door, a VRML scene is no scene, a MacOS or
Windos desktop is no desktop, they are all just happen to be metaphors
suited to explain to the normal user what is going on. But what if one sets
out the go beyond the 'real' world references, creating a spatial language
for cyberspace, based on it's distinct pro and cons? It like the transition
from eclectic classisism to modernism in the start of this century - new
technology, economy and societies will demand new forms and architectures.
>For instance, because of the historical accident of Colonialism, all
>of the world is now familiar with the basic concepts of Western-style
>architecture. It wasn't always that way; in fact, a lot of the
>conquest of various Colonialists was about conquering the native
>conventions on how to set up a landscape. Can't have those messy
>round huts grouped in clusters with little twisty paths between them
>now, can we? It wouldn't bode well for the Cartesian world view....
>Cheers,
>Brandon
Well, let's hope colonisation isn't necessary to build a common code-pool
this time :)
[Thanks for these comments, really useful for my thesis work]
-----------------------------------------
Torbjoern Caspersen casper@due.unit.no
http://www.stud.unit.no/~casper/
Student of Architecture
at the Norwegian faculty of technology, NTH, Trondheim.