Josh

I added some items, and it should work with the IE9 beta. It looks like the beta still has some problems before it is ready, though.

I believe you could make an MMO if you are really dedicated. It may take longer than you originally thought, but you could do it.

The dedicated server renderer is very simple to add, the way LE3 is structured. I basically just have to declare a lot of classes and functions that don't do anything, and it just plugs into the engine the same way the OpenGL renderer does. I like feature requests, because it is easiest if I account for everything I plan on supporting from the very beginning. It's much harder if I try to add features later that require major design changes.

It's pretty easy to do this: Create a buffer. Render to it. Draw the buffer's color texture on the screen.

This is the most impressive thing I have seen you do yet. The detail and variety are fantastic. I mean, this is just amazing for all real-time graphics.

I recommend raknet for the network code for an MMORPG.

Shader class...done. Surface class...done (pretty much). One change in LE3 is the addition of .shader files. This is just a text file with a list of shader components: vertsource="mesh.vert" fragsource="mesh.frag" Your material file then just has to reference a single file: shader="mesh.shader" Of course, you'll be able to just drag the shader file onto the material shader property in the editor to assign it. You will never have to use Notepad, for anything. Now working on "solids". This is based on my old CSG code from 3D World Studio. It's kind of nice that knowledge I developed a pretty long time ago is still so useful. Solids in the engine are a mathematical construct. They're basically a list of planes, with some extra fancy bits that do useful things like test if a point or ray intersects a solid convex volume. It can even detect the intersection of two convex volumes, which is some pretty advanced and fun geometry. Not 100% whether CSG modeling will be supported in the LE3 editor, but there is a pretty good case for it. I would probably program the more complicated CSG brush stuff in BlitzMax, and just convert it to a dumb mesh by the time it gets into the engine, for visible geometry. Solids are used for the visible camera volume (frustum), and can also be used for invisible trigger volumes, blocks of water, etc. I am leaning towards allowing attachment of any number of solids to an entity, and then when the solid is entered by another entity, a Lua script function (or C callback [or both]) would be triggered. This allows you to attach multiple convex solids to an entity to form a concave shape, and the action can then be handled through the regular entity script system. I haven't gotten into it yet, but the entity script system will allow any number of scripts to be attached to an entity, and they will be called in sequence. Special entities will have script functions available that have not existed before. For example, a particle emitter entity will have a script function EmitParticle() which allows you to control exactly where a newly emitted particle appears. Well, writing code for invisible mathematical intersection tests isn't too exciting, but it formed a lot of the basis of LE2, and will for LE3 as well. Then it's on to the OpenGL camera class, so I can actually get some 3D graphics on the screen. I'm mostly just retreading things I've already done right now, but having the code in C++ is good. It gives us cross-platform support for C++ programmers, the biggest segment of LE coders. It allows easier implementation of external libraries. Hiring outside help for some tasks is actually possible now. Finally, with a well-designed C++ core, it ensures we are forever moving forward, and not backing up and redoing things again. If I need to support the XBox or iPhone, that can be done without a massive rewrite or a new language. We learned a lot from LE2 about what works well and what can be improved. I am looking forward to a long and stable future using the LE3 engine design. I'd like to get into game production of our own titles as well, using LE3. I'm enjoying Amnesia: The Dark Descent this week. All I need is an hour or two each week, so when I start a new game it usually takes a couple of months before I finish it. I also discovered the web series The Guild, and found it hilarious, although I am an outsider to the whole MMORPG thing. Maybe Felicia Day will make an episode about my life. Well, I want to have something to show off for GDC 2011, so I'd better get back to work. Enjoy this episode of "Civil Protection". You can rarely go wrong with Half-Life 2 machinima:

Your graphics card cannot run Leadwerks Engine.

There are some constants in the shader that can be adjusted.

With max settings, that's not surprising. Try turning off SSAO and godrays.

What graphics card do you have?

See attached. oalinst.zip

Yes, you can choose from something like 26 EAX presets.

It's already supported. The hardware support for this seems pretty shoddy. Regarding MP3 licensing: http://mp3licensing.com/royalty/software.html

OpenAL is usually already installed on most PCs.

Gravity affects a feather per unit mass the same as it affects an anvil per unit mass. Air resistance would be another force vector opposite of the direction of motion. I believe a surface air resistance force like I described would produce realistic motion.

Interesting question. I guess it would require air resistance proportional to the downwards-facing surface area at any given time.

What do you think?:

I added a few items and tried it out with IE9. Seems to work.

Occlusion culling is enabled by default for lights and animated meshes. The occlusion culling test itself has a performance cost, and it is likely to slow down drawing of other objects more than it saves. You can enable occlusion culling for any individual high-poly mesh, but you typically don't need to.

Hi, I am a pirate. I might or might not purchase your software at some unspecified point in the future. Please give me technical support now.

Frictional Games is the developer of puzzle horror games like Penumbra and the new Amnesia: The Dark Descent. We sat down with Frictional Games co-founder Thomas Grip to learn more about their development process. Josh: I'm a big fan of the Penumbra series. What can we look forward to in Amnesia: The Dark Descent? Thomas Grip: There is a lot to of new stuff in the game. First of all the game is a more streamlined experience both in terms of general gameplay and interaction. For the gameplay we have made sure that the focus is always on atmosphere and story, and made sure that the player is never stuck at a puzzle for too long and stuff like that. One of the ways we do this is to allow multiple solutions so it always feels intuitive to the player. In terms of interaction, it is much easier to interact now. In Penumbra it could be a pain to open doors at times because you needed to click on the right spot and so on. In Amnesia all that is fixed and it is now possible to easily slam shut a door in the face of an enemy and things like that. There is also a new insanity system, a fresh environment, and new story. Josh: Can you talk about the process of developing an idea and turning it into a full-fledged game? Do you start with a design document or just start writing code? How closely does your final game resemble what you originally imagined? Thomas Grip: This time we had a very long iteration period where we tested out various ideas. We actually first had a bite-sized horror design inspired by Super Mario. It was intended to have small self-contained levels and to be played in shorter bursts. It did not fall out too well and we tried tons of stuff before settling where the final game is. In the early iterations we started out with a "fun" mechanic and tried to base a game on that, but as I said, it did not turn out good. Instead we started focusing on the themes and general feel of the game. Then we started building minimal mechanics to support that. To do this we took a lot from Penumbra, as we knew how it worked, and refined it. Once we had the general feel going 1/3 or so is drawn up on paper and we started making maps based on that, then as maps are made we continue to design the remaining 2/3 on paper taking cues from the maps being built. These written paper designs are often very close to the full game and that is a way for us to be able to plan ahead very accurately. Early on we knew around 90% of the assets that we needed, which is great when outsourcing. Then in the final steps the maps get polished and we play them over and over and try to build up the feel we are after. If something does not work out, we often scrap it instead of rebuilding it. Josh: Towards the end of the first Penumbra game, you communicated with a major character over an intercom system. The character never actually appeared on screen, but the emotional connection was established using voice acting. This struck me as a good design decision, because you were able to eliminate all the trouble of making an animated character speak in a convincing cut scene, without detracting from the player's experience. Are there any other examples like that where you made design decisions that helped streamline production, so you could focus on the core gameplay? Thomas Grip: There is a lot of that stuff in Amnesia. For example, there is one character without a lower jaw, which removed the need to do lip-sync. When it comes to visions we only use sound (without models, like in System Shock 2), which in turn also gives the game a more spooky feel. I think having these kinds of constraints also forces you to use your imagination more and not just follow the design of other games. Although it is a bit annoying not being able to do certain stuff, it often turns out positive for us. Josh: What's next for Frictional Games? Thomas Grip: Rest! And tech support... And then we have to see how sales and so go! We have an interesting idea we would like to pursue, but it depends a bit on how we do financially. Check out Amnesia: The Dark Descent for creepy thrills and inspiring game design.

The geometry 3D World Studio outputs is pretty low-poly, so you are unlikely to need occlusion culling within a single structure.

I did not change anything.

The water is drawn its own world in LE2, so yeah, that might be made be a bit easier.