History of Computing [olpc-software] (was AbiWord, HIG)

Alan Kay alan.kay at squeakland.org
Sun Apr 2 02:00:47 UTC 2006 

I think that the only feasible way to do the equivalent of the "Mol Bio of 
the Cell" is on the net, and this would require quite a bit of work to 
organize.

Of course, this was one of Engelbart's powerful goals, to boost the 
collective IQ of adults, especially groups of adults.

Google isn't really set up to do this "Mol Bio of the Computer" but here's 
what I got from just typing "Engelbart" into Google today.

<http://www.bootstrap.org/>Bootstrap Institute: About BI - 11:28am
Doug Engelbart invented groupware. What's he up to now?
www.bootstrap.org/ - 24k - 
<http://72.14.203.104/search?q=cache:4awqRDhD96gJ:www.bootstrap.org/+engelbart&hl=en&gl=us&ct=clnk&cd=1>Cached 
- 
<http://www.google.com/search?hl=en&lr=&safe=off&q=related:www.bootstrap.org/>Similar 
pages - 
<http://www.google.com/search?hl=en&q=engelbart&btnG=Google+Search#>Remove 
result

<http://www.bootstrap.org/chronicle/chronicle.html>Bootstrap Institute: 
Engelbart<http://www.bootstrap.org/chronicle/chronicle.html> biography
At Engelbart's headquarters, his Bootstrap Institute.
www.bootstrap.org/chronicle/chronicle.html - 40k - 
<http://72.14.203.104/search?q=cache:NOpaBBpWuEYJ:www.bootstrap.org/chronicle/chronicle.html+engelbart&hl=en&gl=us&ct=clnk&cd=2>Cached 
- 
<http://www.google.com/search?hl=en&lr=&safe=off&q=related:www.bootstrap.org/chronicle/chronicle.html>Similar 
pages - 
<http://www.google.com/search?hl=en&q=engelbart&btnG=Google+Search#>Remove 
result
[ More results from www.bootstrap.org ]

<http://sloan.stanford.edu/mousesite/1968Demo.html>Doug 
Engelbart<http://sloan.stanford.edu/mousesite/1968Demo.html> 1968 Demo - 
11:28am
On December 9, 1968, Douglas C. Engelbart and the group of 17 researchers 
... The original 90-minute video of this event is part of the Engelbart 
Collection ...
sloan.stanford.edu/mousesite/1968Demo.html - 27k - 
<http://72.14.203.104/search?q=cache:wHtoarMhT8MJ:sloan.stanford.edu/mousesite/1968Demo.html+engelbart&hl=en&gl=us&ct=clnk&cd=3>Cached 
- 
<http://www.google.com/search?hl=en&lr=&safe=off&q=related:sloan.stanford.edu/mousesite/1968Demo.html>Similar 
pages - 
<http://www.google.com/search?hl=en&q=engelbart&btnG=Google+Search#>Remove 
result

The first entry takes you to Doug's site. The third entry takes you to the 
great demo (I was there and its still the "mother of all demos" that I've 
seen.)

So, as far as Engelbart is concerned, it is pretty easy to find out what we 
was trying to do.

BTW, the answer to "how did Engelbart think about the mouse-use?" is found 
here also. He realized that browsing through hyperspace is like dynamic 
information retrieval -- and the main thing you get with information 
retrieval is stuff you don't want! So you need to move onward as quickly as 
possible.

They quickly decided that they had to solve two main things (a) how to get 
sub-second response, and (b) how to avoid constantly moving the hands back 
and forth from the keyboard to the mouse. The solution to (b) was to 
introduce a five fingered chord keyboard for the off-hand.
HermanMillerEngelbartNLS.jpg

The five keys on the left plus two of the mouse keys provided 127 commands, 
which included typing -- the other mouse button (the one on the far right 
in the picture) was CommandAccept. This sounds tough, but the way you 
learned it was to first type commands (like MP for move paragraph). It took 
about 10 hours of use to get reasonably fluent.

By the way, this ensemble was designed by them with Herman Miller. They 
noticed that people liked to scoot around their offices in their office 
chairs, and designed the above carrier that was attached to the chairs. So 
you could lean back, scoot around, etc., while still hacking.

(But you can imagine, just as if the violin or bicycle were invented today, 
this would be rejected because it has a small learning curve, and was going 
into the 80s.)

However, the Engelbartians, and visitors like me, wound up using NLS for 
many hours every day, and thus an interface that permitted great fluency 
made a lot of sense. This idea was also rejected over the years ("no one 
will use a personal computer, and if they did, it would only be for a few 
minutes per day").

The terrific NLS programmers were also able to achieve sub-second response 
on their time-shared SDS-940 (192K bytes, 1MHz) with 17 users. Why? Because 
they really wanted to! It is quite weird for me to use my PC laptop today 
that has more than a factor of 2000 more memory and cycles and not get 
sub-second response. Why? Because most programmers don't understand why it 
is so important for many tasks in the UI, especially retrieval.

So the theory of the NLS interface should now be more clear. The idea was 
to navigate and do short typing tasks with hands apart, and only move hands 
together for typing longer paragraphs. This worked very well. The NLSers 
could type about 30 words per minute hands apart and get regular faster 
fluent typing speeds with hands together.

We liked many things about this system, and there were a few things we 
thought could be improved. When PARC was set up, about 8 of Engelbart's 
folks were part of the original team.

At PARC we added bit-map graphics, overlapping windows and unlimited 
desktops, made our interfaces iconic and modeless (select first then 
command), and came up with a few more ideas about metamedia and simulation. 
Here's what an Alto looked like at PARC:
SmallAlto.jpg

Notice the chord keyboard and mouse! The screen here was the lowest 
pixelation that we did on the first Alto in 1973 (808x606), just enough for 
1 complete WYSIWYG page. As soon as we could a few years later we doubled 
this to 808x1212.

What went to Apple in 1980, seven years later, was pretty much only the 
part of the UI that we used with children (icons, windows, point and click) 
but none of the workflow ideas or the facilities for high fluency. That's 
pretty much still the case for Apple and MS. For example, we used unlimited 
desktops as a way to provide continuity over days and weeks for many 
parallel projects. The desktops were the gathering place for all the 
resources one needed for a project. And, of course, nothing like the 
weakness of powerpoint was needed because the desktops could be sorted and 
organized for presentations, with, of course, the full power of the system 
always available, etc.

I should also mention here another great 60s UI (GRAIL at RAND Corp) that 
was a huge influence. It had a continuous stroke character recognizer that 
was much nicer and smaller than Graffiti (done in 1966), and a really great 
idea of how to do pen based interfaces. There are movies of this on the web.
SmallGRAIL copy.jpg

This was actually the UI that we thought would be the grandparent of the 
Dynabook interface (not Engelbart's). And we did a pretty nice 
object-oriented version in Smalltalk at PARC. The problem was that tablets 
in those days were really expensive. The RAND tablet (the first great 
tablet, invented to do GRAIL) cost $18,000 in the 60s (almost $100K in 
today's dollars). The tablets you could see through in the 70s were about 
$7K (the Alto cost about $20K w/o a tablet). So the mouse, which originally 
was just two pots, won out. I didn't like it because you couldn't draw well 
with it (I could draw and I wanted kids to draw).

Because the character recognizer was so good (you did have to spend about 
15-30 minutes learning to use it), it was possible to also use the 
Engelbart idea of navigation+short-text-input as one mode of use, and heavy 
typing on a keyboard as the other. This worked really well (and would work 
well today).

By the way, I got all of these examples via Google on the net in just a few 
minutes. So some of the great stuff from the past is available in an 
understandable and accessible form today.

Bottom line: some of the opinions about Engelbart and the mouse as 
expressed in this list are curious, since the smallest amount of actual 
delving would provide a definite answer rather than just an offhand but 
needlessly assertive guess.

Cheers,

Alan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/olpc-software/attachments/20060401/71882dbc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HermanMillerEngelbartNLS.jpg
Type: image/jpeg
Size: 18781 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/olpc-software/attachments/20060401/71882dbc/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SmallAlto.jpg
Type: image/jpeg
Size: 20003 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/olpc-software/attachments/20060401/71882dbc/attachment-0001.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SmallGRAIL copy.jpg
Type: image/jpeg
Size: 16186 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/olpc-software/attachments/20060401/71882dbc/attachment-0002.jpg>

More information about the olpc-software mailing list