Gregory Farrington: Don't let Steve escape. He's shown us all
the wonderful tools, but not exactly how to turn it into a way
of teaching science and engineering and medicine easily in one
long weekend. Jim.
Audience Member: Let's take your number that processing
speed goes up ten times every three years for the same amount
of money, do you have any comparable number for how much it costs
to use ten times as much processing speed after three years?
Does it require the same back office of programming and support?
Does it require more to do the fancy stuff? Is there any curve
Stephen Heibein: That's a great question. Generally what you're
going to find is that, I think it's Gate's Law: no matter how
fast Intel makes the chips, Bill comes up with more software
to take up the available nips. I think that happens in all companies.
What we're finding is we're getting more sophisticated software
available on there. And if you take it to the scientific applications,
whether you're talking computational chemistry, fluid dynamics,
those type of applications, what they're doing is they're getting
more robust in terms of now somebody who understands the science
of computational chemistry doesn't have to understand the programming
of computational chemistry, because what they get to do now is
do combinant chemistry and be able to try new molecules that they
hadn't tried before because they're no longer having to do the
programming because the CPU's are that much faster to be able
to do more of the work I guess is a good way to put it.
Now, on the other types of things, the desk top applications,
that's a whole different thing. You're going to find those keep
outdating themselves so quickly, whether it's your talking an
Excel, a Word. Just look at the '97 suite of office is so different
from '95. I can vouch for that. My whole goal is I would love
to see everybody using Web-based tools, which I think people
have found to be more intuitive when they start using them. Right
now there's only a few suites out there that are in Java that
can go across several platforms.
Audience Member: Have you got any results from the effects
of the cyber-education courses, because unfortunately the Ed.
biz has been full of people claiming first programmed instruction
and then CD ROM's, and you know some of the early hyper-text experiments
show that they were actually less effective in teaching than traditional
printed matter. So we're all sort of looking for somebody who
will have an evaluated experiment.
Heibein: No. That's really a problem no-one has. What
was mentioned earlier is how little money is spent on research
in education. I think too often the budgets being such, things
get implemented and are sort of thrown out there and hoped to
be, "O.K., let's find out empirically did it work or didn't
it work. Do we redo it?" No-one's actually tested or gotten
any good data that I'm aware of from any of these. Any other
questions? Yes sir.
Audience Member: Could you provide some more examples
of the development of your intranet services for your corporation
and how it's being used and how far into the organization is it
Heibein: Oh, O.K. In our case, it's quite deep. We have
electronic requisition on-line, on Web pages. You want to purchase
something? You purchase it. It goes in. It knows who your boss
is or what your level of authorization is. It then routes everything
automatically. It's all done on the Web. And what we did is
we took that and set that-we already had a order processing system,
so what we did is we put a Web front end on it. So everybody
interacts with the Web front end, but we didn't change our fundamental
way we do our business on the other side. We just have it so
that users or consumers have a Web look at it. Our vendors have
a Web look at it. They have one from outside the firewall. So
everybody's coming in that way. And I think that's pretty fundamental.
Things like our expense supports are obviously on-line. Everything
we do with HR is not on-line. And what we do is we have everybody
was sent a secure password, mailed to their home, outside of electronic
format. So this way they got that at home, and so you know to
try and maintain security. There's, like I said, it's half a
right now half a million URL's we have inside of Silicon Graphics.
And you look at that, that's a huge amount. All of our projects
are on-line. I can look up any project that is a funded project
with Silicon Graphics, check the status on-line, check everybody
who belongs to that. Generally there will be a link right to
their home page, and I'll be able to do that. My schedule's on-line,
so is my on-line calendar. We schedule our conferences on-line.
Everything is all part of that Web.
Audience Member: I just wanted to suggest that the previous
question goes past a very important point, which is not just packaged
curriculum, but we use Silicon Graphics visualizations programs
to radically increase the number of real time experiments that
undergraduate students do. Like all of them do them all the time
on the ecological territory around our university, doing work
for the county, for the state, and for private business. And
our faculty is absolutely persuaded that is deepens and quickens
the learning by huge factors. And so it isn't all about prepackaged
curriculum. It is about the use of this and other similar types
of technologies as very, very powerful interactive, real-time
learning tools that achieve the very things you're talking about
in a business setting. And I think to miss that misses in fact
the real power.
Heibein: Yeah. Thank you. I think dry experiments are
interesting because now you can try things out in a, you know,
dry environment, take a look at it from different points of view.
I mean over at UCLA, what they did was, after the riots they
wanted to show what it could look like when they rebuilt South
Central. And so to be able to have them understand what it would
look like, they said, "O.K. Yeah. We can put out plans.
We can do little models. Or we can do a virtual walk-through
and have the people who live in South Central see what it looks
like." And the level, the quality that you can now do today-I
mean we can have the drapes look like drapes. We can have the
carpet look like carpet. Wood grain look like wood grain. And
be able to go through things, 30 frames a second, interactively.
And that was really good. And that was very compelling, because
even a lot of architects can't read plans and visualize what it
is just from blueprints. What you want to do is you want to make
sure you can bring this knowledge to everybody, no matter what
level of education. Anything else?
Audience Member: Is this practical for real people? Silicon
Graphics has lots of money to spend on developing absolutely spectacular
educational modules, but Williams College is full of ordinary
professors without-extraordinary professors, of course, extraordinary,
wonderful professors, but never-the-less professors who are not
sitting at Silicon Graphics with enormous budgets for educational
development. So, given the tolls, which are spectacular, how
does a professor who comes in and thinks he might want to teach
physics a different way and have an impact on how his department
functions with students actually proceed, or it hopeless?
Heibein: O.K. Well the, to just reiterate the question
so you have it on the record there: how can this be done by mere
mortals-right?-who need to teach every day and don't have big
budgets, etc.? I think it can be done. There are still some
tools that are being developed. I'm working on one right now that
allows you to go from any-I don't care whether it's Harvard Graphics,
Power Point-you name it. If you've got it in any electronic form
already, it converts it and actually puts it into that format,
puts it in the framework, interactively asks you like, "Do
you have any questions?" "Yes" "O.K., what
are your questions?" And you just type in the test right
there. And it's all done via Web page that creates a Web page.
You never type in one bit of html. We've got it working right
now in prototype, and it's probably going to be marketed sometime
this summer, and going to be a relatively low cost. But it's
one of those things to make it so that you do what you do best:
develop the curriculum. Somewhere along the line you had to
write it down. I don't care what format it is. As long you can
print it we're going to have you print it to the Web, and then
we'll put the frames and all the cute, all the buttons and the
navigation tolls round it for you. So that's where I believe
it's going is to make these things so that anybody can do it.
My boss can do it. I do it, but mind you that's kind of what
I do for a living is technology, so I get into this thing, but
we have administrators doing this. So it's everybody within our
company, regardless of technology level.
Farrington: I guess the implication is if administrators
can do it, anyone can do it. One final question.
Heibein: Yes sir.
Audience Member: I'm a hard-core 3-D graphics technologist,
none-the-less I want to rain on your parade a little bit. The
question really is about tools and authoring environments, and
I think you're reciting the easy part of authoring. Throwing
Power Point slides on the Web is not the problem. If you're talking
about putting together interesting 3-D simulations and animations,
that is an unbelievable amount of work. Putting together a good
multi-media presentation is ten times the work than doing an ordinary
text book which is itself ten times to a hundred times more work
than putting out a set of course notes. I think, as a community,
we are still being ridiculously optimistic about the amount of
work it's going to take to put a really good curriculum, even
an individual course, together using these wonderful new tools.
Delivery is not the problem. Authoring is.
Heibein: I agree with you, especially about the 3-D aspect
of it. 3-D is not, as you noted, 3-D is not intuitively obvious
to the casual user-how's that?-in terms of trying to create something.
I think the things that we're finding that are being implemented
first are the flatware. Basically someone taking something from
Power Point, someone taking things that are in movies already,
bringing video on-line. We're finding a lot of people that are
doing video on demand. What they'll do is they'll have from a
server, especially within a campus community where they can sort
of regulate how much bandwidth they have to the classroom from
this server, those are the things that have been working real
well. Inside corporate walls, like you said, going back to Mariott
example, they'll do from their headquarters or from regional headquarters
right to individual hotels so they can do training there. Now
those are the things that are getting implemented first.
That 3-D is-pretty much what you're going to find is a second
and third wave of people who started doing development. Once
they get excited, they can do things a little bit more challenging.
What I do believe a good way to go though is, when you start
looking at things that are for 3-D, is using tools that are doing
data mining and data visualization. So you might have something-I
might be a sociologist. I might be working with SAS. And to take
SAS and to capture the 3-D output of something like that and use
that and make that so that people can use a slider bar to move
up or move down the graph on that based on being able to go in
and just alter it that way. That's probably the best way to go,
and actually I have an example of that, but I'm out of time.
Farrington: I think that's a good ending. A very good
point, especially because it points up something that I think
is absolutely true, and that is, for all the chips in the computers,
it's the authoring that is very expensive. And if you're going
to give us any tools, and revolutionary tools, make revolutionary
tools that makes it possible for a regular professor to have an
impact on his class and on his teaching relatively easily in a
way that doesn't appear insurmountably difficult or expensive.
That's-that's the gap at the moment it seems to me. Thanks very
much to both of the first two speakers.
We have a slight schedule change. You'll see it from the schedule
that you have in your pack. Originally John Sealy Brown was going
to be with us. He telephoned on Friday at 5 o'clock to say he
caught a Malaysian bug. I don't know what Malaysian bugs are
like, but presumably he's experiencing it, and he's experiencing
it in Palo Alto. So instead of John Sealy Brown, unfortunately,
we're going to have a break right now until 11:00, and at 11:00
we'll reconvene with what we call "The Future Student".