Dynabook Usability
Alan Kay
Alan.Kay at squeakland.org
Sat Aug 23 19:00:25 UTC 2003
There have been various studies, but not covering all five of your
questions as stated.
1. The clamshell idea appears in a number of early drawings of mine
in studies trying to figure out how to make somewhat larger portables
-- one of them was briefcase-sized, some were smaller. It has also
been invented several times independently (as far as I can tell). In
this phase I just made a survey of things that people were willing to
carry and then tried to see what kind of computer could be fit into
them, and when it could be done.
A summary of the early Dynabook studies reveal that it should weigh
no more than 2 lbs and be quite thin in one dimension (so it and
other things could be carried in one hand). A number of different
calculations from 1968 onwards indicated that about 1M pixels were
needed to do "mostly everything". This number remained quite constant
over the years and approaches (only being able to contemplate 500K
pixels or less in the early days of PARC was one of the irritants
that led to the overlapping window idea).
I still like the basic Dynabook form because it is straightforward in
many nice ways.
There were two other basic forms for the Dynabook (which, after all,
is 95% a SW UI concept). One was that we thought there was a better
earlier chance for a display on a chip that could be introduced into
ordinary glasses to make a completely inobtrusive HMD. This was
inspired by Ivan Sutherland's first virtual HMD that was built at
Utah (but was anything by inobtrusive at the time). The other form
that was talked about for the Dynabook was Nicholas Negroponte's
notion that we would soon be living in a world in which wireless
networks, sensors and displays would be part of every room -- and he
realized that great things could be done if your (say) wristwatch was
also a transponder that knew where you were and also knew the
position and orientation of your hand in space. This coupled with
voice recognition would provide incredible opportunities for great
UI. His ARCMAC group at MIT built some of the 70s greatest UI designs
and demos around these ideas. Needless to say Nicholas' ideas have
been "rediscovered" several times since, usually without giving him
credit.
2. At Utah around 1967, a "universal display" was built that had a
nice table in which you could plug in any or all of the input devices
then around. These included Engelbart's mouse, the RAND tablet, many
different kinds of joysticks (including strain gauge ones),
trackballs, etc. Only two devices really worked well: the mouse and
the tablet, and the tablet was best for almost everything. The mouse
had two good features: it stayed where you left it, so it was easy to
find again; and its relative movement allowed small movement stroking
rather than long reaches. The latter was easy to fix with the tablet
by allowing a changable preference for relative or absolute sensing.
The former was fixed many years later in the "tabmouse" stuff we did
at Atari and Apple.
Later, numerous folks, including Tom Moran and Stu Card at PARC
did "Fitts Law" studies on many aspects of user interface and their
input studies confirmed the Utah results in an interesting way: that
only the mouse and the tablet were "Fitts Law" devices and acted as a
fluent extension of the hand. (BTW, both studies showed that the
strain-gauge nonmovable joystick (as introduced much later by IBM and
unfortunately still put on many laptops) was a very bad pointing and
drawing device (one of the worst in most respects)).
3. The GRAIL recognizer at RAND in the late 60s was very good
(generally better and more flexible than Grafitti) so there existed
considerable early experience with recognition. Also, Engelbart had a
non-keyboard input and navigation scheme using a mouse and a chord
keyboard. Both of these worked very well. The conclusion was that a
perfect recognizer or perfectly fluent chorder wasn't fast enough for
paragraph input, and that a much better UI would involve two styles
of use: (a) nav & limited input of text off the keyboard, and (b)
using a regular keyboard (maybe with the Dvorak layout -- or perhaps
even using a stenotype keyboard and recognizer (this was studed
also). (BTW, an on-screen keyboard with tapping is quite a bit less
efficient than a decent character recognizer.)
The most important feature seemed to be that once you chose one
style, it was good if the UI permitted you to stay there, and
limiting "back and forths" was very good. The Engelbart scheme was
very good at this. With "hands apart" you could navigate, issue as
many as 3 commands per second, do limited typing a correction of
typos at about a max of 30 wpm, and generally zoom around. For bulk
typing you would move both hands to the regular keyboard where you
could type at 90 wpm or more.
For the Dynabook, where I wanted a stylus for many reasons, I
adopted this general scheme of nav&rec with the stylus and keyboard
for bulk input. Interestingly we couldn't get the Mac folks to value
this well enough to adopt it for the Macs -- and both the Mac and
Windows today require considerable back and forth movements .... And
we couldn't over-ride the marketing folks at Apple enough to get a
real keyboard on the Newton.
4. I think you would need to make a real Dynabook (or close to it) to
do this test. Aesthetically, I find the Transnote to be a really ugly
and bulky concept. It doesn't pass any of my other criteria...
5. This is a matter of pixels and acuity. The (in Japan) Sharp Zaurus
with a full 640*480 screen at around 200 pitch is a lot more usable
than the 1/4 VGA displays common over here. But absolute size does
count for all ages. My favorite paper medium (until we can do a real
Dynabook) is the Strathmore Sketch Book, which folds to 5.5" * 8.5"
and opens to provide a writing and drawing surface that is 11" * 8.5"
(in landscape) and for which my measurements and scans of the stuff I
put in there would be adequately handled with about 1M pixels (that
calculation again) and antialiasing. We could make one of these today
with two of the smaller 8" diagonal SVGA displays that are now
available pretty inexpensively. It would have a bit of a line/gap
running down the center, but so does my paper sketchbook.
Cheers,
Alan
At 9:32 AM -0700 8/23/03, David Faught wrote:
>Through the history of the Dynabook idea, have there been studies done
>of the ergonomics of:
>
>1. the Dynabook form, versus say a modern notebook PC, or a tablet PC?
>
>2. a mouse versus a pen/tablet?
>
>3. a keyboard versus an on-screen keyboard or handwriting recognition?
>
>4. the flat design of the Dynabook versus a folding design like an IBM
>Transnote?
>
>5. the visual standard paper size (almost) of the Dynabook versus a
>small form PDA?
>
>__________________________________
>Do you Yahoo!?
>Yahoo! SiteBuilder - Free, easy-to-use web site design software
>http://sitebuilder.yahoo.com
--
More information about the Squeak-dev
mailing list
|