2023 November Update
🕑 Added 2023-11-03 07:02:07 +0000 UTCGreetings everyone!
October is already out the window, and November is here. Time is moving way too fast, and not given me enough time on a computer. Story of my 30s... In any event, we've made progress on several things across all products.
First off, FBS Milestone #3 housekeeping. We raised $737.98 since the last update. This puts our total to $1584.28. Milestone #3 may be funded before Milestone #2 is complete. If that's the case, we'll bundle them together. Remember, Milestone #3 will be a programming milestone, so features, features, features. It's what most of you want. :)
Speaking of Milestone #2, work has been somewhat sporadic. When I reach a point in the new UI where I can hand it off to Serhey, I then have to wait a week for him to hand me back a sample file, which, in turn, he has to wait for me to hand it back to him, then I have to wait for him, and so forth. The issues compound by neither of us being on the computer all day like we used to, for various reasons.
Nonetheless, we have made progress. While the new UI for the designer isn't polished, I have finished the core functionality. We've also successfully implemented the entire replacement of model components in the new designer. In the old designer, a vehicle was a single model, and you attached other models on top of it. Morphs affected the vertexes of the entire body design. In this new system, a vehicle comprises many different models, and you can switch out these models seamlessly. For example, we may have 5 different Fender designs for you to choose from, each with its own set of morphs, its own materials, etc.
Here is a small sample video of replacing portions of a test model in the new system.
https://www.youtube.com/watch?v=THtcmcQYiQw
Please note this is not the final product! This is just to give you an idea of what I am talking about.
I've also created a new file format to handle both the UI and linking the model information together. The new file format also includes a tool in the mod tools, so once we get everything set up, art bounties will be less hands-off for me in the future, thus allowing me to put more time into AeroMogul while still generating content for 2nd Gear.
We're improving vehicle materials by adding shader support to the vehicle models. This will allow Serhey to add details to the vehicle designs and improve the quality of the vehicle paint. Shaders have been a problem with this engine in the past. It spent a little while fumbling with getting GLSL to work. I have abandoned that path for now and opted to implement Cg shaders into the 64-bit builds. I use Cg shaders in the 32-bit build, but I omitted them from the 64-bit builds. Why? Cg shaders are a bit of a regression, as Nvidia discontinued Cg shaders 11 years ago. But if it still works, it works, right? It should be fine until we add an Apple Metal or Vulkan renderer. At that point, we'll have to rewrite the shaders, but we'll have to do that anyway since they use different languages than DirectX and OpenGL.
Here are a couple examples of the new shaders which we'll be using on cars. These again are test textures. The real thing will be coming soon.
Now I have passed the ball back to Serhey and I'll wait for more artwork or issues/technical stuff I need to do with the engine.
AeroMogul's progress has been off and on over the last month. When I am waiting for Serhey to pass updates along to me, I am either doing busy work with GearCity or working on AeroMogul. Since the last update, I've done more work on the passenger processing loop. As I added more data, the old loop became wildly inefficient. To counter that, I flatted out the pixel-to-pixel demand into a single array.
This change uses more RAM, but the first-leg calculations have improved to 2.5 seconds, down from 34 seconds with two legs in the old system. The major downside to this change is that I will need to rewrite the second leg system. And the second leg won't simulate Consumer Behavior as realistically as the old system. Storing multi-legs in the flat array would increase RAM use astronomically. Calculating the second leg during the first leg loop would result in a similar slow performance. Luckily, I have a couple of ways I can solve it, but I have put it on the back burner for now so I can move the game along beyond working on the inner game loop.
Moving the game along, AeroMogul is now generating seating, baggage, and amenity revenues. The server sends this data to the client, where the client displays that information when selecting a route from the world map. The next time I can touch the AM code after I get my barrings straight, I'll be working more information into the UI when selecting a route. After that, I have a list of other features to add.
Unfortunately, that's where I am with AeroMogul. I've bogged down with GearCity/FBS, AM's core loop code, and real life. For now, I'll skip efficiencies in the loop code and push to flesh out the game more.
