Channel 7, Plan 9

Gregory Alan Bolcer (
Sun, 22 Feb 1998 20:41:40 -0800

So it's Sunday and they were made for rambling. I don't
like getting into the object fray, but love to watch, from an
academic viewpoint, of course. It's either that or watch
the final episode of Shaka Zulu on channel 13 on my
roommate's television.

WARNING: I didn't want to, but somehow this turned into
an ego-surfing trip; either that or a ton of needless tangents.[ASIDE 1]

> Compact format
> for network transmission? I'm not trying to say Java is fantastic
> and first-of-a-kind. Just that an "interpretive langauge + network"
> does not equal "the equivalent of Java".

Before Java came out, I was a firm Python believer. Python was and
is a fantastic, first-of-a-kind interpretive language with network support
and cross platform capabilities. Having built
a 'mixed' system in C++/Ada/Python which included interlanguage
RPC ala Q, all lovingly co-mingling with a really nice model-view-
controller presentation architecture, I knew all the hacks. As
we started to port the system to other platforms, we ran smack
dab into cross platform, diverse networking, concurrency,
and protability issues. The Python inner circle stopped talking
to me circa 1994q1 when I started questioning their cross
platform applicability when I couldn't even compile the 'virtual
machine' across Solaris 2.1 and SunOS4.1. It was supposed
to be a shortcoming of my technical skills and not thier code. [ASIDE 2]

When I first saw Java it was Spring 1995. Sriram Shankar introduced
it to me. It was beautiful; it addressed all of the problems I was having
with Python and then some. As a user from alpha2 to the present 1.2b, let
me tell you, I've fought every single cross-platform bug imaginable in Java.
They have most of the issues worked out by 1.1.3, despite the spurious
issues of file URL composition and local paths. I was a firm believer
in Java back then, but was still a little skeptical about this WWW thing. [ASIDE 3]

> Wow. Would FoRKers agree with that? The nerve of this guy,
> suggesting we have no "society" :-).
I think he meant no "class". Rohit has lot's of class(es); I've spent
my whole graduate career ditching them. You end up doing a lot of
work for very little credit. [ASIDE 4]

> This is ridiculous. Let's agree on some basics here.
> 1. Model representation, independent of rendered view, is a
> Good Thing (TM). Object/Java/MVC folks know this. So do

I was curious about what the state fo the art of this was. I go
to AltaVista and see what's happening in the world of MVC. I put
in "Model View Controller Java" and the very first link is some
comments that I made a long time ago trying to splain how you
might do it in Java.

That's the problem with the Web, you can't get rid of anything.
David Jones of the Hypertext Bibliography Project has a reference
to my TOCHI paper on event-based, MVC work. Even though it's
not refernced explicitly, can I claim I've made it since he's
at an MIT lab?

Even odder still, I found out that I was a UIST'96 reviewer.
I am deathly afraid that I was supposed to read a paper that
got swept off my desk and is now buried in on my floor.
Some poor bastard to this day is probably pissed that he got
a non-returned review and he had to re-submit it to HICS.

Incidentally, Adam if you read this? Did you ever do a television
interview for your 'Lead or Leave' stuff? I remember clearly
seeing someone speaking on this on the 'talk show' circuit way
back when.

> Hmm. There's a lot of truth in the above. But it's not always true.
> Money
> & innovation *can* mix, they just don't necessarily always mix.

Speaking of which, an open question to all you web-heads. Fame or
fortune? Was there ever a point that you thought the Web was just
and interesting prototype and were not sure of it's success at all?
When NCSA and later Netscape became the primary drivers of the commercial
side of the WWW did you feel like you were losing control? Did you feel
like there was enough competing interest that it didn't matter? What was the
final straw that made you decide to open up the technology and not
form a company? When did you decide it was no longer a prototype
and ready for prime-time?

Inquiring minds...


p.s. That's it, I'm all out of stories. I might have a few
more about machine learning, but that was a whole nother lifetime.
Apologies to those who have heard them before. Maybe one or two more that
are still playing themselves out. I am in a full crush groove coasting
from my 30th last November.

p.p.s. Capirinhas have honey in them too. I make them by heating honey
with water, and then pouring it over ice along with fresh squeezed

[ASIDE 1]= Internet Surfing
Interesting to me, about December '93 I scanned and posted my
surfing pictures from 10 years of exploring Mexico and beyond. At
first, it generated a little interest as it was something other than
physics, CIA World Factbook, and Info & Computer science. Later
it became overwhelmingly popular as it showed up as the first
link on AltaVista any time someone typed in 'internet surfing'. About
April '94 Ron Britvich ( (old?)) and I talked about
putting live video feeds up on the WWW of famous spots all over the
world. I actually had a camera set up in HB at some point in time and
had discussed getting VC from my friend's business associate and
legendary surfer, technophile, eccentric, and owner of the absolute
most fascinating surfboard collection 'Flippy Hoffman', whose
daughter Joyce won the 1966(?) Women's world championship of surfing
and was in 'Endless Summer'.

[ASIDE 2]= FileNet
Late in Fall, '93 and again a year later in Fall, '94, a group of us
went over to FileNet Corporation in Costa Mesa. For those of you
who don't know, FileNet owned 80% of the workflow market before the
advent of the WWW. We went over there in Fall '93 the first time and showed
them a prototype about a 'visual' workflow language we had developed.
They paid their $15k IRUS dues, and said, that's interesting, but
we really don't care unless it runs on PCs. We spend a year revamping
the system, improving it, putting the clients on Windows and Mac
platforms. The clients are in C++ and Ada, the workflow descriptions
and behaviors are in Python. Cross platform cooperaton, now we're
cooking! We even work out a way that you can download the Python
'Handlers' with the client making calls to an HTTP server no less.
We give a second demonstration in Summer to Fall of '94. The review?

"What is this HTTP stuff? Are you going to require that everyone
has that on their desk? We dont' need that stuff. Thanks, again,
but what do you guys want from us?"

They go on to steal the look and feel from the tool, start a new product
line, but nobody really understands the importance of cross-platform,
mobile workflows. One of the graphical workflows from our demo shows
up in PC Week a year later. We write them off as idiots, and my
suspicions are confirmed when their founder Ted Smith sells 100%
of his stock in the company and their price takes a dive from
around $50 to about $14. I am somehow convinced that they
could have turned their corner on the workflow market into
a corner on the intranet market, but then, that's probably
neither here nor there.

[ASIDE 3]= Sun Labs
I gave a talk at the ARPA Workship on Process Technology
in May 1995. I am very ashamed of this talk as it
was my first 'major' one I had given on my own with
no supportive audience members. I had completely botched the
presentation and worse, Fedex had completely destroyed both my
monitor and Sparc 20 that I had shipped out for the demo. I ended
the talk after 15 minutes with 15 more to go and ended up speaking
to my 'insert demo here' slide fielding questions about HotJava/Java
stuff and what our experience with it was to date. I fly back
and on the next day fly up to Sun Labs and give the same 15 minute talk
sans demo. I ask about what they were doing with Java there, blank
looks all around. I ask hoping to get to meet the developers
and discuss some of our issues we encountered "Java was developed
here right?" to which the reply is that it must have been Sun Labs
Europe. Nobody in Sun Labs knew what Java was.

[ASIDE 4]= Estrin
I had a multimedia networks class in Spring '93. My project
was experimenting with tricks for sending images, sound, and video over
the Internet. I was supposed to determine if HTTP was a good transport
mechanism. Working full time, trading commodities full time,
and going to school full time, It was too much work, so I dropped the class. 8-)