This week was spent experimenting with vehicles but mostly the unfortunate task of re-converting various game models, mostly the vehicles. The new engine update that restored vehicle functionality required changing the format of the physical models. A time consuming process when you have a lot of models that require them.
We're using the vehicle test program to quickly let us inspect the vehicle models, test weight and wheel parameters and materials. Some good observations came out of it. Some not so good.
The screen-shot right shows the children's charity truck, or "We Care" pick-up, crashing into a stack of large explosive barrels. Accidents can and do happen.
It allowed me to experiment with animating GUI elements. The slide-out drop-down gadget is styled in black and yellow. Themed after the classic Longbow 2 and now adopted by the new Operation Flashpoint Dragon Rising.
The vehicle selector lets you quickly load up any of the vehicles found, assuming they are built with the appropriate wheels and working physics model you can take it for a test drive. Any problems are quickly apparent.
Mission editing, and how to assemble the data necessary for AI behaviour is now at the stage where I'm building GUI elements which will make the transfer into the main project. Key to rapid building of missions from a players point of view, coincides with rapid development of content from our point of view. Rather than build missions in a seperate editor or off a menu option. Stealing a great idea from the recent Golden Joystick 2009 winner "Little Big Planet", you will be able to pause the simulation, flip into "editor mode" make changes and switch back into simulation mode. The editor is part of the game, the game is part of the editor.
No loading or wait times while experimenting with object placement or trigger logic. The next week will be more work on UI and editor elements. The terrain system doesn't easily allow for decals but I managed a rudimentary terrain hugging poly which only glitches (z-fighting) on small bumps on the terrain as the resolution of the disc and terrain don't match up. There are shapes that deal with this better. I feel a bit stupid in that, in typical programmer fashion, I spent two hours on a code solution to this, when I could have easily made shapes in a 3D editor :) and used the grass shader which contours the mesh to the terrain.
Live and learn.
I'm not sure how live-editing will work out in multi-player. Player objects have states as do all non-static objects for recording and playback. Being able to "rewind" states to make edits to a scene is an offshoot of the requirement to send them over the network and provide VCR like playback. It can be a lot of data but not all of it is critical. In multi-player, it could get badly out of sync. Back in 1996, with Longbow 2, editing a mission was problematic, sync was a real issue over dial-ups. Broadband lets us get away with a lot. We'll just have to wait and see how it plays out. But if it doesn't work well in multi-player it will have to go as a feature. There's nothing stopping you from mission editing in solo play and connecting up when edits are complete. There's only about three thousand steps between here and there. So I'll worry about it later :) My brain tends to work holistically, seeing everything at once so it's hard not to think about these things in relation to any one aspect.
Until later.
No comments:
Post a Comment