Re: Comments Sought - Browser as launcher revisited...
Re: Comments Sought - Browser as launcher revisited...
- Subject: Re: Comments Sought - Browser as launcher revisited...
- From: Ric Phillips <email@hidden>
- Date: Thu, 04 Jul 2002 15:46:39 +1000
I have had quite a bit of experience using "Konqueror" on Linux systems,
which merges the web-browser GUI with a hierarchical file system windows
explorer like interface.
I have my doubts that there is a stand-out model for file-system access that
offers easier useability than the others.
File-systems are arranged as nTrees that can grow far beyond the limits of
user-memory. The number of possible mappings between any 'place' the user is
at, and the places a target 'might' be, reach absurdly large values very
quickly.
The real problem - and one I do not think can be overcome by passive design
techniques - stems from some immutable factors;
1) Human content classification is not bound to (logical) storage
structures.
2) People don't have one classification system for anything. We use
multiple, overlapping and extendable classification systems that will never
map onto the object-attribute structure of file-systems.
3) Computer file-systems have a very high degree of recursive similarity -
and the objects they contain are also highly similar.
4) Things in the file systems can move about a lot. And even a folder icon
moving it's position within a window is a significant change from the point
of view of navigation effort.
Point 3 and 4 are very important because all mammals (including us) evolved
to use pretty much two complimentary navigation strategies. A) Orientation
by easily differentiated and immovable landmarks, eg, the bend in the river,
the three pointy mountains on the western horizon, the large rock that looks
a bit like a wombat, etc. B) Remembered kinaesthetic passage through
perceptually similar environments (related to the ubiquity of fractal
patterns in nature), eg, walk for a while between those trees towards the
three pointy mountains, turn my body to this particular side and progress to
where the grassland starts, then just follow that line of dirt (track) in
the grass that has been made by animals going down to the river. (Oddly
enough research has shown on average, male mammals prefer the landmark
strategy and females the local knowledge strategy - rotate the rat-maze in
your lab, and it's the males who become disoriented!)
Other research has shown that the more points of difference their are
between the items that make up any collection of objects, the larger the
number of objects in that collection whose location can be recalled without
effort or 'mnemonics'.
File-systems do not offer sufficiently, obvious and stable 'landmarks'.
Their objects are too similar, and offer to few perceptual points of
differentiation (Which is why Simpsons and Peanuts cartoon icon sets that
have no relation to content classifications actually help people navigate).
Navigation uses a kinaesthetic 'metaphor' - there is insufficient
proprioceptive differentiation in a series of mouse clicks or keystrokes to
support instinctive local-knowledge strategies.
Put all this together and a passive interface is never going to cut the
mustard.
Within 5 to 10 years however, a combination of adaptable software agents,
and speech recognition will have buried the 'file tree' model, as deeply as
the the physical organisation of files on your hard drive is now buried
under that same logical tree.
Ric Phillips
Faculty Web Coordinator
Faculty of Humanities and Social Sciences
Latrobe University
Room HU3 324
Phone: 9479 2792
-----------------------------------------
_______________________________________________
applescript-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/applescript-users
Do not post admin requests to the list. They will be ignored.