scary_pigeon

I'm fed up of working part time of Jet Thunder, so I'm going to quit my job and work on JT full time to make sure it gets done in reasonable time. currently refactoring code to aid the development of AI and Physics. Jet Thunders code currently mixes the graphics objects with physics data. By separating the aircraft physics into a separate static library I believe jet thunders game code will become more agnostic of whatever physics engine we use (currently Newton) I like Newton in many ways. Despite being the work of apparently one man - It doesn't fall over. Is well documented compared to Bullet, but not the fastest. For a number of reasons it would be nice to switch to PhysX. with this code refactoring, this becomes a possibility. And with the physics at last decoupled from the graphics objects, then driving the physics on a separate thread should be more straightforward. Our framerate is pretty good - but its always nice to have more. we remain, mainly single core at the moment. some important fix recently. JT is now far more likely to run on other peoples PC's than just ours - as developers we have PC's with dev software on - but because we're not a studio, we don't really have naked test PC's. Dante discovered on some PC's at his work place that JT wasn't running and I was able to replicate the error and fix. This demonstrates the necessity of open beta testing at the end of the project to catch these dumb errors. But one wonders: the publisher and investor types we showed JT in the past, probably didn't have dev software installed on their PC. It perhaps made us look poor if our game just crashed without comment. At least this shouldn't happen now.

updating the flight model - its long needed some love. Hopefully in a few hours I will enjoy flying the skyhawk in the physics test shell. previously, the flight model was using complex equations. Now we're going to make better use of interpolation and pull the flight model back into C++ rather than lua script. This will give us code efficiency boost. Especially with lots of aircraft present in the battlespace.

we're still working on this project. recently been reworking the terrain engine to be less art driven, in the sense of the artist having to laboriously present everything: a more pipeliney approach. so we can now do better layering and tree/hedgerow placement. don't worry about the glitches on this screenshot.

there will be some new screenshots coming soon. we now have much better looking sea thanks to shaders. :)

the purpose of this is to let you know that even when there are not new movies and screenshots - things are still being done! this last week and a bit - has basically been allocated to making the system much more modable - certainly for the team. previously the engine has been more of a test shell - with this source code, I hard code the spawn position of the aircraft and the other objects in the scene, then we test the Flight Model - test other things like guns, bombs, AI - stuff like that. Even the aircraft was hardcoded and full of emerging legacy from other test shells. The Pucara class had transmuted into skyhawk behaviour and appearence and all filenames were weird - mirage this, mirage that, etc. so we've been reorganising the files and making it more general - to work towards a fully script based approach of setting up the scenes/missions. Other things that need doing include scripts to place weopon pilons. this was another hard coded test with the bomb. That is mostly done. More recently I've working on improving the athestics of the damage model. My favourate flightsim is IL2:FB. It was always very satifying to knock off a wing with the cannons. This is what we want to see and do in the Jet Thunder series. It will possibly take a few more days to do. - hopefully we will be able to then make another little movie and screenshots to show this occuring. Next we'll probably be working on basic carrier take off and landings. With that we should have basically enough for an argentine carrier lauched raid on a british warship mission. -- but shortly - im off to my job at walmart. A dissapointment today. I missed the courier who had my new GFX card.

we're in the stage of seeing what works. I'm writing the underlying battlespace sim now to support the realtime gameplay to give depth to our missions. Not sure what is best design. is it where you fly till you die? or is it best where you die and its like, so what - i'll just fly another plane (until there are no more planes left). I prefer this sort of campaign where you interact directly through possession of an aircraft on a per mission basis. we'll try a few different things to see what works best. The design though aspires towards a design missions and move units about possibility.

and svn repository is versioning system - lets us developers ensure we have coherent versions on each others machines- we can update the central repository then everybody else gets the changes when they update. whats been worked on most recently is UI system, working to move away from hard coded keymaps so users can customise that aspect. Got to fix a host of bugs, then we continue to polish the demo.

most recently upgraded the ground physics - and also making the new ground physics work with ships rather than just terrain. It is again possible to sit on an aircraft carrier with a skyhawk or harrier. also doing house keeping task - setting up new SVN repository for the developers.

still working towards beta - we have what we imagine would pass as an alpha version with many elements of single player gameplay in - but buggy, glitchy. beta testing ought to be when the game is free of all known bugs and were all gameplay features are complete. maybe. certain game modes might be beta testable soon. How long things take is dependent mainly on how much time we work on it. We imagine that Jet Thunder may help us found a studio when its finished, but until then we have our own jobs. Was making great progress when the last company I worked for got rid of half its workforce, me included :-) but I got a new job :-/ so is back to evenings and weekends. at least this company is relaxed about me working on side projects.

yep hoping for end of 2009.

matrox sent some screenshots back of Jet Thunder demo operating on their systems. thought some of you might be interested - I just wish I had one of these toys. http://www.matrox.com/graphics/en/products/gxm/th2go/

17 foot beats my 32" :)

It is now January 09 - it has been an interested month or so. A couple of months ago I got made redundant, so inbetween that job and my current one, have been able to advance Jet Thunder quite rapidly. Its now in a sort of demoable state, no user interface, just lua scripting that can define everything - setting up missions and triggering events. But there are things that cannot be shown yet such as the way the surface to air missiles AI is not well done yet. Missiles will loop around the plane until they hit or exhaust themselves rather than if missing, passing onwards. Some of the aircraft AI is not ready to show to general public either - when flying in a straight line, the plane rocks left and right too much - like when Maverick's mate is spooked and can't land his plane. The good thing is that the main gameplay features like the different sorts of AI is there - just that they need fixing and polish - Also, much improved is the ground physics. which reminds me, the new ground physics only works with terrain - it will fall through other objects at the moment so I must fix that. At this stage in the project there is a lot of bugs and issues like that. So after the AI is fixed, and after a user interface system is written, then I think the project enters content period where we recreate historic and what if missions using the scripting system and possibly some tool that we might develop... ...and then there is also dynamic campaign and online play options to work on. So a lot has been done, but theres still a fair bit to do.

yeah i feel the same way :P - though a widescreen TV at 1080p is not too bad. Will have to implement trackir soon. I think trackIR and triplehead2go for 3 monitors would be good gaming experience. I think the only head movement you might need is up and down as you'd have such a good horizontal view.

got some inner stuff working now. game play related AI stuff, ways to move from one AI type to another - this is needed to build missions.

bomber AI now - i had some fun with windows movie maker... http://uk.youtube.com/watch?v=QgeXadiQ9mI

ai landing plane on carrier. doing bombing ai next.

am being made redundant from my previous company - therefore free again to work on this great project.

its as dante says. :)

wow, most impressive.

we're all familiar with headshake. great flightsims like IL2 have it. it seems like a good idea to give an impression of what its like to experience manouvers in a plane. what nobody has is eye roll. If you get a video camera and hook it up to a tv - see how crazy and disorientating it is when you roll the camera 45 degrees one way or the other - the whole world tips on its side on the TV... ...yet if you on right now roll your head 45 degree - things done seem to roll like they on the TV. why? what is the difference? the difference is eyeroll. if you look in the mirror and put your head up 90 degrees one way or the other the eye actually rolls within the socket to maintain a global alignment with the vertical - this is a balance cue. so I have been wondering if the virtual eye in a game should align to the current gravity field example: this is how things I think would look from the human eye if doing a gentle turn. this is how I think things would look from either a fixed camera or a human eye responding to gforcing which tend to make the pilot feel as though the G field is through the bottom of the plane rather than down to the ground as in a very gentle bank. this will be a turn on and offable feature. Personally I think in tests it is absolutely wonderful - the missing link to my real experiences of flight and being on the ground playing a sim. Dante though who has also been up and about in the air thinks the effect is awful - disorientating and completely false- so it is clearly a taste issue. it gives you some advantages - when zoomed in from the cockpit just looking at the target - you can often become disorientated - if alls you can see is earth or blue sky you have no visual feedback of your roll angle. but seeing the the target on the hud and the hud glass roll because of the eye makes things like straff attacks when diving easier (for me). I am currious what other people think though.

be interesting to move along with that. currently i just kill time working on old JT stuff. maybe i opensource it. from what i've heared its surprisingly good this TSE thing and all i've done may as well be binned :) - though would be porting cockpit code and developing capaign stuff. cross platformness could be sweet.

yeah i'm aware of it too. and it surprised me.

i feel as though i must appologise for slow progress. i've done scarse amouts of work on the project in the last month and a bit. this is for two reasons. I recently got a new full time job and there is a sort of shock to the system of trying to find time for jet thunder and beer. and when i got over the shock, well we're gamers as well as developers - and i've been addicted to dawn of war: winter assault. did some JT coding today though. some server code.

super :) cheers.