| Q&A |
 Jon
Peddie,
President of Jon Peddie Research (JPR)
26/08/04 |
| Q1 |
A
brief presentation of Jon Peddie Research. |
| A1 |
Jon
Peddie Research (JPR) monitors, forecasts and consults on the
graphics and multimedia markets and advises the leading companies
on corporate, technology, marketing and product strategies.
We also perform due diligence and “sanity” checks
for investors and senior managers of companies. JPR maintains
elaborate databases and files on graphics and multimedia, components,
markets, companies and other related subjects.
JPR
offers its consulting clients a bi-weekly technical marketing
advisory analyst’s report, Jon Peddie’s Tech Watch,
and a quarterly report on market shares of the semiconductor
and board companies, as well as various special market reports
from time to time.
The
combination of extensive hands-on experience, comprehensive
databases, interviewing and reporting, and actual testing
makes JPR uniquely well qualified to help companies and investors
in the market. JPR can provide a unique perspective and knowledge
base of the market, its dynamics, trends and players. We monitor
the market everyday all day—this is our only business.
The
staff of JPR is a group of industry specialists with over
55 person-years experience in computers and computer graphics.
One of the pioneers of the graphics industry, Dr. Jon Peddie,
started his career in computer graphics in 1962. After the
successful launch of several graphics manufacturing companies,
Peddie began JPA in 1984, and after leaving JPA in 2001 formed
Jon Peddie Research to provide comprehensive and customer-intimate
data, information and management expertise to the computer
graphics industry. Peddie has taught courses on display and
image processing boards and lectures at numerous conferences
on topics pertaining to graphics technology and the emerging
trends in digital media technology. Recently voted one of
the most influential industry analysts, he is frequently quoted
in trade and business publications and has published over
70 articles on computer graphics as well as appearing on CNN
and TechTV. Peddie is also the author of several books including
Graphics User Interfaces and Graphics Standards, High Resolution
Graphics Display Systems, and Multimedia and Graphics Controllers.
|
| |
|
| Q2 |
You
have been involved in computer graphics since 1962, what is
the more significant event in the CG industry since this date? |
| A2 |
That's
a tough one, there have been so many advances in hardware and
software. If you look at what we could do in 1962 the comparison
to today is truly amazing. In '62 we had stroke writer displays,
and computers that were built with magnetic core memory and
integrated circuits that had only simply flip flops. Graphics
were only line drawings and were either created with tedious
assembly language or arcane Fortran. Only the elite had computer
graphics, the primary applications being CAD and weapons simulation.
The systems were time-shared (one main computer, several terminals),
there was (at least one) operator of the main computer plus
tech support for the terminal users. The main computer was as
large as four to six refrigerators, and the terminal looked
like giant radar screens on a special table. (I may have some
pictures of this stuff if it would be helpful and/or interesting)
The screens were monochromatic (white or green), had simple
text (dot matrix style) and only lines for graphics, no images,
no fill. A main computer costs $100,000 to $250,000 (in 1962
dollars - about $2 million in today's money) The ARPANET (forerunner
to the Internet) wasn't developed yet and data was exchanged
over distance by a stack of punch cards or a reel of magnetic
tape. Conversational graphics consoles were developed by General
Motors (DAC-1) and MIT Lincoln Laboratories (Sketchpad), resulting
in computer-aided design (CAD). Sketchpad used the first light-pen,
developed by Ivan Sutherland, and the IBM 7094 (the first commercial
computer to use semiconductors) dominated scientific computing
at the time; and researchers in England created the first computer
with virtual memory. This computer was capable of 200 KFLOPS,
your mobile phone is about 100 times more powerful.
Today,
on a $1,500 portable PC we can create cinematic quality graphics
and produce images that are so lifelike they have to be studied
to be certain they are not a photograph. That has come about
due to the semiconductor revolution that is epitomized by
Moore's law, the explosion in sales of PCs and associated
devices, the shrinkage in size and price while quadrupling
in capacity of disks, the abundance of cheap memory, common
high-level programming languages like C++, and the deployment
of common user interfaces like Windows. Programs and data
travel at near the speed of light from and to anywhere in
the world, and even our grandmothers can enjoy high quality
computer graphics in the form of games and email animations.
Now
of that what is the most significant CG event? Assuming all
of the infrastructure was a given (i.e., we have low cost
highly powerful computers with high-level programming languages,
plenty of memory, and great displays) then one thing that
gave CG a big kick was texture mapping. It spawned, or enabled
the spawning of mipmapping, anisotropic filtering, bi and
tri-linear filtering, full-scene anti-aliasing, and new exploitations
of alpha plane techniques. If you asked the graphics chip
producers the question they would, without hesitation, say
programmable shaders. So I guess the answer is two-fold: texture
mapping and shaders.
|
| |
|
| Q3 |
3d
contents are still very rare on the web today. How do you explain
that fact? |
| A3 |
That's
a lot easier; three reasons: limited bandwidth, no efficient
programming model, and lack of a universal user interface. |
| |
|
| Q4 |
We
are celebrating the 10th anniversary of VRML... but, is VRML
dead? |
| A4 |
We
refer to the people who practice the art of VRML as furry vermals.
VRML was developed to make moving CAD files easier. Some people
with great visions, (perhaps too much) imagination, and seemingly
unlimited energies pushed VRML into the realm of virtual reality
and from there into web 3D. It never satisfied any of those
visions and a lot of hearts and wallets were broken. VRML still
isn't dead, but then neither is Fortran. Things don't seem to
die in this industry. But neither Fortran or VRML are mainstream
or considered as a viable vehicle for anything any more. And
I know this will bring out of the wood work and from under the
rocks all those VRML lovers who still toil away at proving what
magnificent thing VRML is, or will be with the next release
that is. |
| |
|
| Q5 |
3D
Chip makers predict the end of pre-render 3D in few years. What
would it change? |
| A5 |
Real
time rendering already exists in several areas - you experience
it every time you play a modern first person shooter like Fry
Cry, or Half Life. In the large studios where elaborate compositing
is done, rendering still takes longer and still relies on very
large very powerful machines and takes days to render a few
minutes of the film. So the goal is to drive that to real time
and it will happen, you could almost predict the exact date
if you froze the level of special effects. For example, if you
took some movie (e.g., Spiderman II) and took five minutes of
a scene of it and used it as a benchmark, and said (for example)
that it took four HP or sgi workstations four days to do it,
then with Moore's law you could say, by 20xx it will only take
4 seconds. The problem is at 20xx we will have come to expect
special effects (that we can't imagine yet) as the norm and
they will still take four HP and sgi workstations four days. |
| |
|
| Q6 |
Is it possible to predict the future of CG? |
| A6 |
Gee,
I thought that's what I just did. :-) |
| |
|
| |
|
 |
|
|
