General development plans

[Josh]

Leadwerks is nice to use, but there's not a huge compelling reason you have to use it. With Ultra Engine, performance is the big benefit it brings. The whole idea of "optimizing" your game goes away, because it's already optimized when you use Ultra. I learned how to do this with all the user projects I tested and inspected with Leadwerks. Then when NASA contacted me, I saw they were using very detailed CAD models in VR. If they were making a game, the same model should have been lower resolution and collapsed into a single mesh, but they wanted to visualize the full-resolution models with full articulation. It was kind of the same problem I saw with indie developers who did not optimize their game content, and the idea for Ultra Engine was born. This is not something that is intuitive to AAA developers because they consider optimizing the game to be an important part of their job, and they feel very proud of their ability to do that. So even John Carmack doesn't really understand this idea the way that I do, because he is accustomed to working with a team of people who can spend large amounts of time on whatever needs to be done.

I can see into the near future that once the new editor is finished, we're going to be in a situation where there are tons of feature requests of all different kinds. Some of those things should be built into the core editor and some of those things should be added as extensions, written both by me and third parties. Leadwerks did not have this ability, with with Ultra Editor you can access the entire editor internals by Lua script. It's extremely cool because you can add new tabs, move things around, and do anything you want. It's like having the source code except you can add things without forking the code and breaking compatibility. As long as the API stays the same, you can add new features without increasing the technical debt.

I don't want to do a 3D model store. It would make more sense to plug into an existing database. CGTrader.com and Turbosquid have a very large selection, but Sketchfab is newer and more reliable, because everything is available in the well-defined glTF format. So at some point it would make sense to add Sketchfab downloads straight into the new editor. Surprisingly, model pack DLCs have always done very well. In fact, I can sell more copies of a DLC than I can give away as a free download, the exact same item. If I find a consistent high-quality source of models then this is something I will consider. I'd prefer not to release 3-4 DLCs, I would rather do zero or 20, but not somewhere in between. Providing you with 3D models is not part of the job description  in my signature, so I should probably not think about this right now.

There will eventually be a need for some kind of repository for user-made extensions and components. Extensions are written in Lua and add new features and tools to the editor. Components are written in C++, Lua, or C#, and add new behavior that can be added to an entity. (I created an entity component system for C++ that uses a precompiler to make the code components modular, but I haven't written about it yet.) I am picturing a system most like the extensions browsing in Visual Studio Code. It doesn't necessarily need to be a commercial thing, just a way to search and install extensions. However, my rule for this is that the content should come before the "store". Once there is a large number of nice extensions available, maybe 100, then the system for browsing and installing them from within the editor comes.

It's still being discussed, but I want the licensing model to force me to pay attention to the users here on this forum. The situation with Steam was kind of strange because the people asking for features on the forum were not the place money was currently coming from. So instead of the active users here, I was trying to make the software appeal to new users on Steam, most of whom bought it on impulse during a sale. It was a strange situation and I felt like I was saying "no" too many times, and going off in weird directions that were annoying to people. I want to just focus on the tools and engine and say "yes" more often to the things that you ask for. Game Launcher and Steam Workshop were created because I was searching for the right direction to go in. I was floundering around looking for a direction. This was before NASA contacted me and I focused in on performance and realized it was possible to create an engine that was really ten times faster.

What I am aiming for is very fast performance, which is particularly useful for VR but great for all games, with a stable long-term product (ten years) built on the new architecture that constantly improves without any backtracking. And I want to not waste times on extraneous distractions that don't really help you, so I recommend you use itch.io for game distribution, Sketchfab for 3D models, etc.

[Josh]

Things that are not my job:

• Running a social gaming website.
• "Helping" you with promotion of your game. I do not have any useful advice on this
• Distribution of your games. itch.io, gamejolt.com, indiedb.com, and Steam do a much better job of this than I can.
• Providing 3D models. Sketchfab, CGTrader.com, and turbosquid.com do a much better job of this than I can. It's better for me to just make sure the glTF loader works perfectly with all files.
• Writing books and academic papers for tech conferences. It makes me feel very cool but doesn't really help you guys.
• Giving you bad life advice or humorous articles that no one except me finds funny.

If I am ever wasting time with something that is not in my job description please let me know.

[Josh]

In the past, I have had a rule of trying three new things each years. Sometimes these would lead to very good outcomes, like publishing Leadwerks on Steam or working on VR projects with NASA. But I think the time where crazy new ideas is needed is over. This year I should try zero new things and just finish the idea all this has led to, which is Ultra Engine.