> 1) Hello Peter, thank you for accepting this interview. Could you please
> tell us who you are, what is your background and you current occupation?

I've been programming for most of my life.  It is, and has always been, 
one of my main hobbies.  I almost always have two or three projects in 
progress at any time.  Art of Illusion is by far the largest and longest 
lasting of them.

As for my background, I have a Ph.D. in Applied Physics.  My graduate 
research involved developing algorithms for simulating protein molecules. 
Since I finished graduate school five years ago, I've been working as a 
programmer, creating software to help biologists analyze certain types of 
data.

> 2) How did the idea of coding your own application come to your mind and
> why did you choose to write it in Java?

While I was in graduate school, I took a class on computer graphics 
algorithms.  I did that just for fun - it had nothing to do with any of my 
research.  I had a great time in the class, and decided that doing 
something related to computer graphics would make a fun project.

As part of that class, we had to work with some truly horrible modelling 
tools.  Figuring there had to be something better, I downloaded demo 
versions of a few commercial modelling programs.  Their user interfaces 
were a bit better, but still pretty bad.  I decided I could make a better 
one.  And so I started writing Art of Illusion!

The choice to write it in Java was an easy one.  Through most of graduate 
school I worked in C, but I started to play around with Java a little bit 
near the end.  There were so many things that were just so much easier to 
do in Java - memory management, threading, etc.  Now that I'm used to 
Java, I would hate to have to go back.  Also, the idea of being able to 
write a program on one operating system and have it "just work" on every 
other was a wonderful one.  Not that Java really achieves that goal - you 
still have to test on every operating system you plan to support and fix 
platform-specific bugs - but it comes closer than any other development 
environment I've worked with.

> 3) Java is known (or has been known) as slow and not very Real Time
> Graphic friendly. What is your point of view on the subject? Did you face
> difficulties because of using Java, and how did you overcome them? (PS:
> please note that I'm not a programmer of any sort, and that I only know
> about this "bad" reputation without knowing if it is deserved)

This is an ancient myth that should have died long ago.  The very first 
versions of Java were slow, since they were entirely interpreted.  But 
modern virtual machines have "just in time" compilers that make it as fast 
as (or faster than) most statically compiled languages.

Of course, every development environment has its own performance 
characteristics, and writing efficient software requires you to understand 
them: what operations are fast, what operations are slow, what types of 
designs that environment is well suited to, and so on.  If you're used to 
writing C++ programs and then switch to Java, your first programs may end 
up being very slow - not because Java is a slower language, but because 
it's a very different language, and requires different techniques.  C++ 
also had the reputation of being very slow when it first came out, and for 
the same reasons as Java: it took time for the compilers to mature and for 
people to learn how to use it properly.

> 4) Could you sum up the history of Art of Illusion, with a briefing of the
> very important moments of its "history"?

That's a harder question than it sounds.  The first version was 0.1.  Each 
version since then has incremented the number by 0.1, and it's now up to 
2.0, with 2.1 due to be released shortly.  So I'm really up to the 21st 
version.  Every release has added major features, so there aren't many 
distinct milestones, just a steady, gradual progression.

> 5) How strong is the community of AoI users? Are they only CG enthusiasts
> or are there some CG pros among them?

That's also a difficult question.  There are probably about a dozen people 
who post regularly on the discussion forums, but many times that who have 
posted just a few times.  I get an average of one or two emails each day 
from new users.  But according to SourceForge, Art of Illusion has been 
downloaded over 130,000 times, with 200-300 downloads every day.  So there 
are a lot of users out there, but I don't know who most of them are.

> 6) Do you pay attention to other Open Sourced 3D applications, like
> Blender or POV-ray, for example? What do you think of them, of their
> advantages and disadvantages?

Not all that much attention.  Every now and then, someone will point out a 
feature they like in another program, so I'll look at it to see whether I 
could add something similar.  But I don't worry about trying to compare 
them.

I think one of the greatest and least appreciated strengths of open source 
software is that it doesn't need to compete.  That's very different from 
commercial software, where you constantly need to keep an eye on your 
competitors so you don't lose sales to them.  First and foremost, I'm 
writing Art of Illusion because it's a lot of fun.  If someone finds it 
useful, that's great.  But if they decide to use a different program 
instead, that doesn't hurt me.  And if somebody tells me they like another 
modeller or another renderer better, my first thought is not, "How can I 
compete with it," but rather, "How can I interoperate with it so users can 
combine the best features of both programs?"

> 1) What is your prefered aspect of Art of Illusion, the one you like the
> most to code and involve yourself in (e.g.: modeling tools, animation
> tools, rendering engine...)? (and obviously, explain why! :D )

All of the above, and many others.  What I really care about is finding 
interesting problems to solve, and just about every aspect of computer 
graphics is filled with interesting problems.  Given my background in 
numerical algorithms, I suppose the most rewarding parts are the ones that 
give me a chance to invent completely new algorithms.  Some examples 
include the subdivision surface modeller, the method for raytracing 
displacement mapped surfaces, and the procedural texture editor.  These 
are also the parts that take the most work, of course, but seeing the 
results of that work is very rewarding.

At the other extreme, I can definitely tell you my least favorite thing to 
work on: the boolean modelling tool.  It was pure agony to write, and it 
took over a year to get most of the bugs out!  I then discovered there was 
a fundamental flaw in the algorithm I had chosen, which causes errors in 
certain situations.  I've been meaning to rewrite the boolean modeller 
with a different algorithm for several years.  But it was so painful the 
first time that I keep putting it off.

> 2) The render engine is quite complete, with lot of very interesting
> features, like Global Illumination, HDRI, photon maps... Do you plan to
> extend photon cast to implement Sub Surface Scattering, for example?
> Generally speaking, what are the plans for the render engine?

Better support for scattering materials is definitely the biggest hole in 
the raytracer right now.  You can expect that to be filled sometime in the 
next year.  The other big change to the raytracer which I plan to make in 
the near future is to parallelize it so it can take advantage of multiple 
processors.  Dual processor computers are becoming common, and within a 
few years it will be hard to buy a computer that only has one processor. 
I want to make sure the raytracer can take advantage of that.

> 3) What are the weaknesses of the tools of Art of Illusion you would like
> to work on in near future? Any modeling tool you plan to add, for example?

Overall, I'm very happy with the modelling tools in Art of Illusion. 
They're mature and very powerful.  On the other hand, I've learned that 
different people prefer different styles of modelling, and you can't 
expect one tool to satisfy everyone.  I designed the modelling tools in 
Art of Illusion to work in a way that is intuitive for me.  Some people 
really like them, while other people would prefer them to work in a 
different way.

The most important recent development in this area is the PolyMesh plugin 
that was recently released.  It provides a completely different type of 
mesh, and allows a very different style of modelling.  Some people who 
aren't comfortable using the standard tools in Art of Illusion like it a 
lot more.

> 4) The animation system of Art of Illusion is quite satisfactory. For
> example, you can do skeletal animations and set morphings according to
> bone positions in order to simulate muscle masses. What are the plans for
> the future?

I think Art of Illusion has reached the point where it provides all of the 
basic tools that are required to make good looking animations, as well as 
a few more advanced tools.  But animation is a huge field, and I could 
spend years just adding new animation features.  For example, there's the 
whole field of physical simulation, something that I've barely touched on 
so far.  Or advanced character animation techniques, where you define not 
just the skeleton for a character but also their muscle structure, and 
then the skin automatically moves realistically as the muscles flex 
underneath.  And then there's particle systems which are useful for many 
different types of animations.  I've written a Particle Jet script which 
can be used for simple particle effects, but there's so much more that 
could be done.

I don't have any specific plans for animation features right now.  At some 
point I'll suddenly feel inspired to write a particle system engine, or a 
collision detection engine, or something like that, and so I'll do it. 
That is how most features get added: I simply do whatever I think will be 
the most fun.

> 5) Do you know of interesting scripts and plugins under development behind
> the scenes?

Very little happens behind the scenes, but if you read the public 
discussion forums you will discover all sorts of things that are being 
written.  A very exciting development over the last year or two is that 
other people have started to write plugins that add really major new 
features to Art of Illusion.  For example, there is the PolyMesh plugin I 
mentioned earlier.  Or a plugin that can render scenes to vector image 
formats like SVG.  Or a very powerful procedural editor for creating 
plants and trees.  It's great to see other people starting their own 
projects to add the features they want.