VRML conference: Java - What is Smalltalk Tomorrow? 
Author Message
 VRML conference: Java - What is Smalltalk Tomorrow?

This is an important note.  Please read it carefully.

I am finally getting around to report on my vrml conference.  I
am putting it here because of the important implications for
Smalltalk.

The VRML conference was about setting standards in the Virtual
Reality Modeling Language.  It is a 3D replacement for web pages.
Microsoft, Sun, Silicon Graphics, Apple, and others were there in
force to put their mark on what happens with 3 dimensional
graphics on the internet.

I saw a demonstration of 3D stuff written in Java that knocked my
eyes out.  It was food for thought and I think it could mean a
great deal to the future of Smalltalk.  I will describe this
after I get some other things out of the way.

First and perhaps most important at the conference I found out
that Java is Bill Joy's baby.  This is important because Bill Joy
is one of the few technoTypes around who is comfortable both in
Technology and Marketing.  He was instrumental to Sun's early
success because of his very broad and complete understanding of
the hardware technologies which were used in workstations and at
the same time had a firm grasp on Marketing aspects of those same
technologies.  He knew where his hardware was coming from and how
much it would cost him years in advance.

In the past I have broadcast information about Bill Joy and Sun
in the Smalltalk community because I thought that someday this
would be of some use.  That time has come.  There are those of
you out there that heard me get up and speak several years ago
about the dynamics of what Bill Joy did for Sun and that the
Smalltalk community should learn something from this experience
and act accordingly.

I hope this gives me a little more credibility when I tell you
that Bill Joy is at it again and what he is doing is Java.  He
predicted the current state of affairs in respect to Java's
position in the market over a year ago, and again he has turned
out to be right on the money.

I suggest that ParcPlace/Digitalk position themselves to take
advantage of Java rather than fight it.  They are doing some
great things with Java and Smalltalk and both lead and follow in
respect to these things.

(Aside: May I also suggest that PP/D needs Dr. Deming's methods
in order to bring about the quality that will be critical to
PP/D's future.).

Now more about the conference.  I saw several things there
including great ideas which will certainly make their way into
Smalltalk.  For example there was a Hyperbolic space VRML
presentation for untangling and making visible the connective
structure of the Internet.  The whole universe out to infinity
was mapped to a circle completely visible on the screen.  This
was done in a manner which could make any arbitrary node "the
center" with all associated nodes subsidiary.  This could be done
in a way so that continuous change from one node to the next
could take place much in the same way one pans in 3D.  (Technolgy
wise the only difference between a pure cartesion view and this
kind of view are constants in a Homogenous co-ordinate
transform).  One could move from any node to any other node in a
continuous way with visual clues never leaving the screen.  The
size of the visual clues were display-related to how many
connections one had to go through to get to the other object.  In
other words, all directly connected or near by objects were all
visible.  In one case they showed where over 100 objects were
directly connected.  Clearly this same technology is due to
become part of Smalltalk, sooner or later.  For example, it would
have great use in viewing Class Hierarchy or connections among
classes.

Also, Diamond Park, the 1 mile square multi-user virtual world
was there.

The best of the show, however, was Michael Deering, one of Bill
Joy's hot shots, showing off 3D stuff done in Java.  Most
impressive was that the quality of the display was equal to or
better than the movie "Toy Story".  The stuff was not canned.  He
was moving around in 3D using "DOOM" like keyboard controls.  He
could apply arbitrary behaviors to objects using controls written
in Java.  This included dynamically changing the polygon
fracturing of a dinosaur in real time, reflections and all.  The
frame rate was actually better than that of a movie.  Generally
THERE WERE NO ARTIFACTS that I could see in his presentation.  

His slide presentation was done in 3D with behaviors of slides,
words, characters, and various graphics.  He "typed" his slides
in 3D and displayed them "star wars" scroll style as well as
arbitrarily flipping them in 3D, changing light reflections off
of the tubular text of his slides as well as embedding them in a
3D space with a scenic road with his various exhibits sitting on
it.  We got to see the exhibits as he moved his view position
along the road. He jokingly showed us an "overview" of his talk
by zooming his camera position back so that all of his slides
were visible on the screen.  

We zoomed into a scene sitting on top of one of the characters of
one of his slides.  He zoomed down until we saw a ruler which was
1 micron long.  He then zoomed out to 1 foot, 1 yard, 1 mile, and
then finally to a view the actual size of the earth.  He said the
limit was light years across.  They were using 250 bits of
precision for the graphics.

I hope I haven't scared anyone about Java.  If you remember back
to the early days of Xerox Parc.  There was a relatively small
group and they did a considerable number of amazing things.  They
did things graphically, by the way, which far outstrip things we
most often seen in Visual Works today.  I suggest that today
graphics needs to get back into Smalltalk because tomorrow will
be too late and opportunities will be lost.

In considering what Smalltalk is and where it ought to go, I
would recommend using the two following books as guides.  I am
seriously tempted to go out and buy every person at
ParcPlace/Digitalk a copy of each of these:

"Out of the Crisis" by W. Edwards Deming
"Built to Last" by James C. Collins and Jerry I. Porras



Mon, 15 Jun 1998 03:00:00 GMT  
 VRML conference: Java - What is Smalltalk Tomorrow?


Quote:

> ...
> This is an important note.  Please read it carefully.

> ...
> I hope this gives me a little more credibility when I tell you
> that Bill Joy is at it again and what he is doing is Java.  He
> predicted the current state of affairs in respect to Java's
> position in the market over a year ago, and again he has turned
> out to be right on the money.
> ...

This is a follow up to drive home an important point.  

HERE IS THE IMPORTANT POINT:  WHAT BILL JOY IS DOING WITH JAVA AS A
STANDARD IS THE SOFTWARE EQUIVALENT OF WHAT SUN DID IN ITS EARLY
DAYS WITH HARDWARE STANDARDS.  So Bill is doing it again...and the
high growth rate is characteristic of both situations.

Why harp on this?  Because there a number of ignored opportunities
for the Smalltalk community (and PP/D) to do this.  All we have to
do is discover them.  If we don't somebody else like Bill Joy will.

This is exactly what I have been trying to tell the Smalltalk
community for years.  The power of a >>>REAL<<< commitment to

Quote:
>>>REAL<<< standards is overpowering when it comes to growth and

position in the industry when correctly applied.

Sun's early growth from nothing to the largest installed base in the
industry required an extremely strong commitment to hardware
standards.  "Oh! you have the fastest, smallest, most economical
hard disk (ethernet, memory chip, CRT, ...) in the industry?  Come
right on in and we will plug you into our next box!".  This is taken
for granted now days, it was not then.  

I used to go around spouting off about how someone ought to do the
same for software, in particular Smalltalk.

Here comes Java.  Well, I told you so.  I have stood up more than
once at OOPSLA conferences and other settings and said we ought to
do this for software. (I can't remember exactly, but I may have
stood up said this from my seat in the audience in the OOPSLA
general session in San Diego in '88 and in Arizona in '91).  Also
some of you may remember me telling in this forum how when I
contracted for IBM they solicited input for their Smalltalk group on
marketing strategy and how I responded with the Sun story (this was
before I knew about Java).  This became part of their marketing
strategy (eg. "Whatever the standards are, that is what we are going
to do.")

MY SUN STORY
Sun's early success was driven by an extremely strong commitment to
standards.  My manager at CTI, Siemens (Georg Eberharder, Now CEO at
CTI) and I went across the highway 101 to Sun's small demonstration
room when they were just getting started and had just made their
first major sell.  It was to Computer Vision.  95% of their
presentation was about standards, standards, and more standards.  
They said in effect that whatever standard emerged, that was exactly
what they were going to do.  About the same period I heard Bill Joy
talk with a breadth of understanding about workstation hardware
technologies.  It was clear that it was a matter of course the he
and Sun were going to do the right things when sticking all these
technologies together in a box.  Compare that to HP who continued to
hook their desk top 9000/500 series machines (a machine that was on
the market BEFORE Sun came on the scene) to a hard disk using HPIB.  
The same time Sun had a hard disk you could carry around in your
hands, HP had hard disks the size of washing machines.  The
difference is that Sun had real standards and HP had introverted HP
standards.  (no negative reflection on HP, every good company makes
mistakes)



Tue, 16 Jun 1998 03:00:00 GMT  
 VRML conference: Java - What is Smalltalk Tomorrow?
This is quite interesting stuff, but I'm not entirely clear on how you
are suggesting to apply it to PPD. Do you have some concrete
suggestions for how to implement this commitment to standards and
quality. It seems to me that this could be interpreted in a very wide
variety of ways, e.g.

- drop Smalltalk and start producing Java environments
- work with Sun to get the Java VM improved and capable of running
Smalltalk as quickly as possible
- make their own web products work well with Java (I think they
already do)
- try to get a Netscape/Java style deal for Smalltalk with some other
hot application.
- generally work harder to get their bugs fixed and keep their
customers happy.

I'm not clear which if any of these you are specifically advocating,
and I'd be interested to know. I'm a firm proponent of the theory that
if you want a company to do something, you need as many customers as
possible pushing them in the same direction at the same time.

Also, a few minor quibbles with the points you raise, although
certainly nothing that invalidates your basic points.

- The standards-orientation in VisualAge considerably predates IBM's
involvement with the product. It's been a feature of Envy/Smalltalk
for quite a long time.

- In touting Sun as an example of how to succeed in this industry
using standards, don't forget that the company currently dominating
the industry has shown at best mild interest in standards they didn't
invent.

- Even Sun's attitudes towards standards and "best-of-breed"
technologies have typically been confined entirely to hardware (see
e.g. OpenWindows and NeWS).

--
 Alan Knight                | The Object People


 +1 613 225 8812            | Ottawa, Canada, K2C 3N2



Tue, 16 Jun 1998 03:00:00 GMT  
 VRML conference: Java - What is Smalltalk Tomorrow?

Quote:

>- Even Sun's attitudes towards standards and "best-of-breed"
>technologies have typically been confined entirely to hardware (see
>e.g. OpenWindows and NeWS).

Perhaps you could clarify what you mean by that, since it seems to contradict
Sun's use and promotion of software technologies and interfaces like BSD, the
Internet protocols, Java, Motif, NFS, Posix, SVR4, XPG, X, and Wabi.
Yes, Sun initially picked some losers too, but that's not a reflection of
their ``attitude'', surely.


Tue, 16 Jun 1998 03:00:00 GMT  
 VRML conference: Java - What is Smalltalk Tomorrow?

Quote:

>- Even Sun's attitudes towards standards and "best-of-breed"
>technologies have typically been confined entirely to hardware (see
>e.g. OpenWindows and NeWS).


Quote:
>Perhaps you could clarify what you mean by that, since it seems to contradict
>Sun's use and promotion of software technologies and interfaces like BSD, the
>Internet protocols, Java, Motif, NFS, Posix, SVR4, XPG, X, and Wabi.


Remember how Sun delayed Unix unification efforts by holding on to its
Open Look GUI rather than signing on to Motif?


Wed, 17 Jun 1998 03:00:00 GMT  
 VRML conference: Java - What is Smalltalk Tomorrow?

Quote:


> Remember how Sun delayed Unix unification efforts by holding on to its
> Open Look GUI rather than signing on to Motif?

YES but they had a much better product and were willing to offer it instead.

Ask how many Sun sites use Motif vs OL today. Less than  you would expect.

SOG



Sat, 20 Jun 1998 03:00:00 GMT  
 
 [ 7 post ] 

 Relevant Pages 

1. VRML in Java vs Java in VRML

2. Southeastern Simulation Conference needs VRML/Java papers

3. Smalltalk and Java Conference in Germany

4. from near VRML to REAL VRML : My first EAI / Java / VRML Humanoid Avatar

5. VRML and the Internet of Tomorrow

6. Hello Smalltalk, bye bye Java (was Sprint: Hello Java, bye bye Smalltalk)

7. (Java success = Smalltalk death) or:[ Java + Smalltalk = Future ]

8. RESUME: will code for hourly rate [11 yrs total exp, Smalltalk, Java, VRML, 3D Multiuser Simulations]

9. IT-VRML: first Italian Vrml conference

10. Java in frame can't get instance of Java EAI + VRML in another frame

11. VRML generated from JAVA, not JAVA script?

12. vrml-java-generated vrml

 

 
Powered by phpBB® Forum Software