I got a email from Yahn Bernier

theboomboom · Posted: Wed Dec 07, 2005 9:34 pm Post subject: I got a email from Yahn Bernier

I'm not sure if you guys would be interrested, but I posted it anyways
Cause i'm cool like that

Oh yeah, in case you were wondering, I am writing an ISearch (a story about your research, not just a paper on the topic) paper on the Source engine, so that's what the interview is all about.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rick,

Gabe forwarded your questions to me to get you answers. Please let me
know if you need any further detail or follow-up information on the
answers below.

Yahn Bernier
Sr. Software Engineer
Valve

----------------------------------------------------------------------

-Why did you choose to code your game with C++?

We are constantly evaluating all available tool options and language
choice is just one of these tool choices. C++ was used because the team
had experience with C and C++ in previous projects, because we could
control the performance characteristics of the compiled code (a problem
you run into with managed code like C# or other schemes like Java), and
because it facilitated the Source engine's (COM-like) component
requirements. Of course, we might change to other tools/languages in
the future as we evaluate what will make us most efficient at producing
the kinds of experiences we are trying to deliver.

-What makes a good game developer?

This is a tough one, since we have developers who are really good in
hard to quantify ways and developers who are good in ways that other
developers on the team are not. I can tell you what we look for,
generally, when hiring programmers. We tend to hire generalists rather
than those with domain-specific knowledge, we look for strong problem
solvers, we look for people who get things done, we look for people who
are self-managers, we look for people who have worked on collaborative
projects in the past, rather than always preferring to work by
themselves. We don't have hard rules about number of years experience,
computer science degrees, etc. We try to get a sense for where we think
people will peak in their career trajectories. We like to hire people
from our MOD community because they have shown an ability to create
something new and ship it to the world.

-What was the most important part of creating the Source engine?

There were a few important things, the most important, in my opinion,
came down to building the right team. Other significant decisions
included knowing what we wanted the products (HL2, CS:S, DoD, etc.) to
be rather than just building up technology for the sake of technology.
Other than that, allowing sufficient time for the good ideas to
percolate throughout the game and the bad/unimportant ideas to fall
away.

-When you were still in the planning and concept phase of the Source
engine project, what did you envision it to be?

I think we described it as the highest quality engine in existence that
would allow us to create the "Best Game Ever".

-Has it become what you envisioned?

That's up to our fans. Inside the company, we're pretty happy with
where we ended up and where it's heading as we continue to extend the
functionality of the engine and improve it.

-Would you consider your engine to be a success or a failure?

I think we would definitely consider it a huge success, not just from a
business and financial point of view, but in specific ways which we
considered requirements, such as support for multiplayer gaming, support
for MODs, integration with the Steam platform.

-If you think it is a success, what made it one?
-If you think it is a failure, how could you improve it?

There is a huge list of improvements that we are working our way
through. I don't think there is any large aspect of the engine that we
would consider having failed. We had some tribulations the first few
hours after launch where we hadn't prepared Steam for the incredible
demand for Half-Life 2. That's probably not the worst problem to have,
but we are always trying to improve the things that need improvement.
Internally, we've never satisfied and we hope that translates into
extremely high quality technology and games for our customers.

-Which part of the engine are you most proud of?

There are a bunch of really amazing advances in the engine such as the
physics, AI, acting support, improvements to the multiplayer networking
layer, and amazing visuals from the renderer. The less visible thing is
that the code for the Source engine is so extensible. We have a lot of
internal work going on adding new features to the engine. This would
have been a lot tougher with our previous technology. There other thing
that I'm proud of is the sheer power we provide to MOD times. We give
out a significant portion of the interesting source code to MOD teams
for them to tinker with.

I think the area where we didn't get as much done as we would have liked
is on the tools side. We had great tools, but we got to a point where
we had to table a bunch of tool work just to get Half-Life 2 done.
We're now going back and investing a bunch of time improving our tools.
This will pay dividends not just to our projects, but to our MOD teams
and to our licensees in the future.

-Which programs/tools did your team use to create Half-Life 2 once you
had the engine set up?
-Would you recommend these programs to mod teams?

The compilation tool we used at the time we shipped Half-Life 2 was MS
Visual C++ 6.0 but we recently transitioned the team to MS Visual Studio
.NET 7.1. A few years ago we moved our asset/code control from MS
Visual Source Safe over to Perforce that was both painful and liberating
at the same time. We continue to use IncrediBuild (a plugin for MSVC
which allows for distributed compilation of C++ code). We have a bevy
of internal tools we wrote including map compilation tools and a set of
utilities for doing distributed processing to speed up compilation times
for maps. We wrote our own facial animation tool, FacePoser, and our
own world-building tool, Hammer. We mainly use XSI for modeling and
animation, but some of the artists use other tools in their day-to-day
work in the art pipeline.
_________________
You do not say that the wheel is fatigued, yet the body is as material as the wheel