4.0127 Direct Manipulation (1/75)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Thu, 24 May 90 16:25:45 EDT

Humanist Discussion Group, Vol. 4, No. 0127. Thursday, 24 May 1990.

Date: Wednesday, 23 May 1990 11:36pm ET
From: "Sheizaf.Rafaeli" <21898MGR@MSU>
Subject: Not really Halio, Disintermediation, Direct Manipulation

Blame Kessler for reigniting this one. I'm just here for the ride...
I'd like to share some (experimental) results and run a short argument
with those HUMANISTS not too blinded with infatuation with their own
preferred machine...

I've been interested in a concept system designers call "Direct
Manipulation". In essence, this is a suggested guideline for those
engaged in making software products. It intends to indicate how
best to make computerized interfaces. Following is an abstract of
a research report, titled "SEMANTICS OVER SYNTAX".

"Direct Manipulation" is a theoretical construct used in
explaining the quality of human-computer interaction. DM is often
invoked to explain the appeal and diffusion of emerging hypermedia
and other computerized communication applications. Building DM
interfaces, it is claimed, should make interfaces as intuitive as
the steering wheel. But DM, as a theoretical construct, lacks formal
explication and validation. Here, I suggest one element: an
interface is "direct manipulation" when it elevates semantics over
syntax. Allow the contents of the task through, even at the cost of ceding
the structure so dear to you. DM as transparency.

I conducted two experiments. Each started with an some conventions
about how the "syntax" of computerized tasks ought to be. In each,
the conventional "syntax" imposes some structure on the semantics.
In one, 101 subjects performed information retrieval tasks, using
combinations of selection mechanisms and menu depth structures. In
the second experiment, 56 subjects queried a database using
different access methods: natural language, menu hierarchies, or
complete content schemas. In both experiments, the DM-based access
method proved superior, contradicting empirically-based
extrapolations or conventional expectations, and supporting the
construct validity of DM. Dependent measures include user
confidence as well as the productivity measures of speed and
accuracy. Direct manipulation or the transparency ideal should
receive more attention in theory and research of computerized
communication, as it is not just a convenient design principle. As
a dimension of describing mediating systems, DM may have profound
use in communication theory.

Now, assuming my methods ARE kosher, (please let's spare the
network), I'd like to pose a couple of questions:

1) My thinking is social-science, communication theory based. I'm
sure this business of battling syntax and semantics has some
treatment in the humanities as well. Could you show me?

2) Several disciplines have discussions of something called
"disintermediation". For example, The Reformation is said to
have disintermediate between people and God. I wish to claim that
a "Direct manipulation" policy is a recipe for "disintermediation"
(remove or reduce the syntax, or mediation. Highlight the
semantics, or mediated). Has "disintermediation" appeared in any
Humanities disciplines?

3) "Direct manipulation" was actually championed first by
Macintosh supporters. Some may claim the Mac interface is DM,
what with all the garbage cans on screen, mice in hands and all.
But it seems that Halio is making the argument that these icons
are all syntax that gets in the way of semantics. IBM, on the
other hand, allows the sheer glory of the semantics to shine,
disintermediatedly. Comments?

Sheizaf Rafaeli
Michigan State University, University of Michigan,
and Hebrew University of Jerusalem,
(currently 21898MGR@MSU)