[net.cog-eng] automatically turning tasks into displays

dir@cbosgd.UUCP (11/10/83)

Many software development projects follow an evolutionary
cycle composed of (1) requirements analysis, (2) high-level 
design spec, (3) detailed designed spec, and (4) user guide.

Somewhere in this cycle the requirements functions (i.e tasks) are 
turned into a user interface spec, typically consisting of displays
and the means of interacting among them.  The process of turning
tasks into displays is perhaps 80% art and 20% science.

Question:  Has anyone found a way to turn the description of a task
directly into the "best" means of presenting that task on a CRT, 
especially on graphics CRTs?  By directly I mean that most requirement
specs are on-line, thus the task descriptions can be formalized without
much problem (like using pseudo-code or structured english), and
then special programs can scan through the formalized task descriptions and
<here's the rub> automatically come up with likely displays, interaction
methods, etc.  Has anyone pursued this idea?  

Another way of thinking about this idea is to build an on-line design 
aid with expertise in the human factors of user interface design.
An "ergonomics work bench" a la writer's work bench?

	Dean Radin 
	AT&T Bell Laboratories - Columbus
	fast mail: cbosgd!dir