Friday March 07, 2014 Development Update – We're Getting THere
I yearn for the days of yore when I could just boot up a game, set up a quick scenario, snap a bunch of action screenshots, and be done with a Friday update in an hour.
Actually, I yearn for the day of even yore-er yore. Seems like it was just yesterday when I could say, hey, I’d like to fly plane X. And a month or so later I’d be flying it. It’s only been, what, 10 years since that was the case. Of course, there’s no going back to that. Even the simplest flying games of today probably can’t do an airplane in a month. I myself can barely stand to look, much less to fly, airplanes from 10 years ago. But I am still a bit sad about all the work that goes into the extra quality. If you were to draw a graph on work hours vs simulation quality, the line grows exponentially.
And while you’re here, close but not quite at the self-imposed 100%, progress is very hard to show.
First of all, here’s where we are with the backer rewards section. We really tried to have it operational last Friday, but internal testing showed that once again we had a logical flaw for larger reward tiers with more than one copy of a stock plane selected. Had to drastically redo the flyable selection this week. Lost all the prettiness and ended up with a painfully simple design, but at least it works.
I guess the page is ready for the world, we just have one final task: to merge the kickstarter and paypal database of backers with the DCS site member list, the hardest part of which is dealing with accounts for backers that do not currently have a digitalcombatsimulator.com user profile.
On the aircraft front, we finally do have a delay on one aircraft model, the P-47 external. We had a long unpleasant saga with trying to get high-quality original blueprints for it last summer. That took longer than anticipated, and what we got was not as thorough as we needed. Once we finally got into animating all the bits and pieces of the aircraft, we’ve realized that our data was not good enough, and that some parts would go out of alignment when moved. We had to go chasing pixels, and start fitting parts back together, animating, measuring, refitting, and trying again. This, again, goes back to my earlier point about standards. In an older game, we’d never care that an aileron slightly clips through the wing when fully deflected. We wouldn’t care that a gear strut does a physically impossible warp a few inches to the side in order to move to the down and locked position. And so with DCS, we have to go and make sure every movement is absolutely perfect, spending more time on these details than on the entire model in an older project.
The 109, our flagship, the pinnacle of our hopes and desires, is still a bit shiny everywhere. Not nearly as bad as two weeks ago, but it still fluctuates. Graphics programmers are slowly but surely moving DCS over to DX11, and it’s a process.
The choice for me is extremely painful - show nothing, or show something that does not look great.
The cockpit programming is complete. All the gauges and switches and systems are working, except, annoyingly from the point of a Friday update, you cannot see it in the cockpit yet. We can see it all in debug mode. However DCS cockpits are done in a peculiar complex way, different from what we’re used to from before, and a programmer cannot simply tell a certain needle to rotate a certain number of degrees. As it is during most of the process, the cockpit mesh is a monolith. In our previous projects, we would simply make each moving part a separate object with a pivot, and the programmer would move that object about that pivot when needed. In DCS, we have to do a lot more work on cockpit animations, doing them manually, and we’re still waiting for them to be complete.
In the meantime, all the gauges gauge, and the switches operate all the systems, the radio talks, the AFN2 guides you to airfields (the only objects at the moment you can guide to), everything works, and this would have been a great time to make a cockpit procedures video, except you can only see all of that happening in debug readouts and not in the cockpit 3D model.
So, since I’ve made too many overly optimistic predictions in the past, I won’t this time. I will only say that it’s very frustrating for me to write these kind of updates, that I really want to have a fully functional plane I can take for a spin and screenshot away, and that we really are tantalizingly close.