From: rnk@sei.cmu.edu (Rick Kazman)
Subject: SCI: Information overload solutions? %-)
Date: 16 Mar 92 15:04:31 GMT
Organization: The Software Engineering Institute



I am involved with a project which is concerned with (among other
things) the design of displays for the instructor-operator station (IOS)
of flight simulators.  These displays are characterized by having
relatively high update rates (at least 10 Hz) and are highly complex
(thousands of pieces of information are collected and displayed).  These
displays involve graphical information (map data, radar, planes, etc.)
and textual information.

I have been thinking of ways of increasing the size of the virtual
display area which the instructor-operator (IO).  In the good old days,
IOS stations were physical panels with lots of switches, dials and
displays of various sorts (gauges, mostly).  These phsyical displays
were easily modifiable (just weld on another bench beside the existing
one) and had the nice property that the human IO could physically
navigate around the stations (which were up to 75 feet long), could see
everything at once and had an approximate mental map of where things
were located relative to his current position.

Right now, things are often done WIMP-style, with workstations running X
and Motif.  Lots of windows can be overlayed, and the notion of physical
navigation among the displays has been replaced by hierarchical screens
and menus.  The screen is just too crowded, information is easily
obscured and navigation among displays is non-intuitive.

I have been looking at ways to put some of the characteristics of the
old displays back into the workstation-based displays.  Rather than
having windows overlapping, I want to have a virtual space of
information which the IO can navigate around, in at least two
dimensions, possibly three.

Some ideas which I've considered: 

    o the "perspective wall" notion which Mackinaly, Robertson and Card
      presented at CHI'91.  All information is presented on a wall which
      the user manipulates by "dragging" sections into focus.  All
      other information on the wall is still present, but "fades" off
      the sides of the screen in simulated prespective.

    o employing a virtual window manager (like vtwm) and presenting
      information as a monolithic 2-D wall which the user can move around 
      (so that pages never overlap, and so the user can form a mental map 
      of the shape of the wall).

    o surrounding the user with a cylinder of displays, and consider the 
      screen to be a "porthole" into display-world, which the user can 
      navigate.  This is essentially the virtual window manager idea in
      3-D.

Can anyone comment on these ideas, or point me to some relevant papers
which I should read?

Thanks.  rick

Rick Kazman (rnk@sei.cmu.edu)  	    Ph: (412) 268-5752      | Disclaimer:
Software Engineering Institute, Carnegie Mellon University  |    Nope
5000 Forbes Ave, Pittsburgh, PA 15213-3890                  |