Lead Developer, Stardock Entertainment
Published on January 14, 2005 By CariElf In GalCiv Journals
The primary thing that I worked on this week was decreasing the loading time when starting a new game. I had discovered last week that the graphics were what was taking up the most time to load. So the first thing I did was to create a background thread to pre-load the models so that they would be cached in system memory when the actual objects were being created. That helped a little, but not enough to make a noticeable difference. Since it was the anomalies that were taking the most time to load, Joe and Jake worked on making them take less time to load. In the meantime, I created another thread that would start creating the stars, planets, etc as soon as you leave the Galaxy setup screen. It goes through and creates all the stars, planets, anomalies, resources, etc, and just waits until the Civs have been setup. It still takes awhile on the gigantic galaxy, but it makes it a lot better. While it can handle people spasmodically switching back and forth from the galaxy wnd, I don't really recommend it. We will probably still put in some kind of loading dialog because we can't create the map objects in a thread (they don't create properly), so there is always going to be some lag time between hitting the next button on the Opponent setup window and the game coming up.

I also fixed it so that we only create the galaxy object once, and only load the data files once. This may change if you load a save game that has different files, but usually you'll be only using one set of data, so there's no need to keep reloading it from disk time you start a new game. This helped with the thread that sets up the galaxy, since all I have to clear now are the stars, planets etc. The next logical step was getting it so that you could start a new game from within a game. This had worked in GC1, but it got broken in GC2 when we changed the influence and fog of war.

I also hooked up some more of the data values on the Planet Management screen, including the pictographs. Right now, the data is still being calculated using the GC1 algorithms, but that's something that we can change without affecting the inferface.

I fixed some other random things like the random event text not always parsing correctly.

We had a meeting today so that Brad could give us feedback on the current build, and now we have a whole new (longer) list of stuff to get done.


Comments
on Jan 14, 2005
Yeah you, keep up the good work.
on Jan 14, 2005
Wish we could do more to help you make this game which sounds like is turning into a game and a half or more of work.
on Jan 15, 2005
have to do Y. You also then have the whole weekend to muse over the problems before the crunch meetings where you need to suggest solutions.

Paul.
on Jan 17, 2005
I don't really have a problem having fun on the weekends. Plus, if we think that we're going to need to work on the weekend, it gives us a chance to do so. I actually didn't work this weekend, but I got a good portion of my stuff done today.
on Jan 18, 2005
I spent my entire weekend in work, thanks to a Friday afternoon meeting!



Paul.