scary_pigeon

just boring stuff today. no sound or graphics. just numbers. In preparation for a more intensive environment I have been adding a new datastructure. A particle mesh. In my scheme, there is a grid of linked lists. and the world is then populated. The idea is to let the engine focus on modelling the physics accurately only in the region the player is interested in - what is within the view range. Beyond that, general methods must be applied. The aim is to create an environment of considerable and apparantly seamless complexity. I think it is essential for a dynamic campaign, if it is to feel as if it is progressing for it to be able to all be happening at the same time in many parts of the game world. You must be able to deviate from the flight plan and find, lo and behold, targets of opportunity - that then respond in defence and through calling for help.

now its time to apply our fancy cockpit to the developing game itself. There are several subprojects experimenting with different sides of the overall project - now we must bring it all together. the specific things i've done in the last day are to relate to simple but until now ommited things like text output. normally we put debug and other stuff out to the console window. now we dont need to. we can write it to the screen. also, there are ways for the script to add text to screen now and so this is used to give feedback on use input. simple things like, "Gear Up", "Gear Down", "flaps 54%" and all that. Also a way to chat online has been worked in. our think our general cockpit system is now very complete and nice and programmable. Soon enough we'll have pucara cockpit. but now, I am with pen and paper working out how to bring all the different bits we've shown you together.

i am excited too. such rapid progress is because I've been off work for two weeks. just have to get some holidays to carry on.

I'm dreading going back to work. I've been off sick the last couple of weeks with mumps - a worrying and unfortunate thing for an adult male to get - and using all time available to program. I think I'll book my remaining holidays off ASAP to keep the momentum going. Aim remains a demo for about march. Yesterday, me and dante continued to worry about the terrain. as this self declared deadline approaches we realise that the place holder terrain system has become a bit too permanent and needs improvement. We await an alternative to slot in to be provided by romano, or an enhancement to the current one I use - with the quickly programmed warts taken out. In the mean time, I added a sort of adaptive clipping zone in the test flight version (we have several version we develop to test different features, one is the cockpit shell where the cockpit parts are being focused on - another is an example of landing and taking off from a carrier in the terrain - another is the mission scripting system (which will be the basis of our dynamic campaign)) - it was nice to see an end to z-fighting - the sytem makes the clipping as narrow as possible which enhances z-buffer resolution over any given area. its good stuff. Also, dante continued to point out the short horizon and showed screenshots of how good wwii online looks in comparison to our rubbish short distance. I thought, well thats odd, because on the map in wwii online you can see that the terrain is never more than 15 km - it is just fogged and clipped. So it hought, I'll try that. and clipped the far plane to 30km added a bit of extra fogging and you're now flying at 20,000 ft and thinking, i can see the whole world - because if you dont clip off the corner of the 'table top' it looks like a table top - and when you do clip off the corners, you always get a nice flat horizon - and yet the skydome gives an impression of earthly roundness. Then satisfied and amused by that, I carried on with the project to script the flight model. Now that is basically done and I await more information on exact flight characteristics. I'm sure this isnt right. in our sea harrier the max turning rate is 22 degrees per second at 250 kts - that seems wrong to me! and so we must tweak it to get it right.

Done some interesting things to day. we worried about the terrain. our terrain looks at the moment quite nice low level, but pants when up on high. very short horizon, and bad things like that. This is moving up the agenda of things to sort out. one possibility might be to redo the terrain system with lessons learnt and speed and horizon distance added. There are pluses and minuses with the approach i use. The system I would develop would be a sectored, load on demand LODed system - which is what we have now - but with lod deactivated. The principle weakness of this approach could be that it might be slightly slower than a more in vogue ROAM algorithm - but it might I think be faster in other ways. The way I did things, was very good with getheight functions which would tend to be demanded more often with a more intensive scene. A strength of the system is that the sectors are loaded and compiled into display lists very rapidly and seems to suffer not too baddly from load stutter. anyway, all thats still a bit on the back burner. what I have been doing today is opening up the flight model. its now being moved to LUA script. This may enable others to step in and makes whats wrong write, and allow us to tweak more easily. It will be a form of prototyping that we may port back into C for speed after we're happy with the behaviour - in anycase - for AI aircraft - the physics will probably not be scripted or if they are, done with simpler calculations. I think we can afford to waste a bit of CPU time getting just one aircraft to fly accurately - the players aircraft. Tommorow, will probably complete the experimental port into lua script, of the harrier FM. with it laid before him i am hoping dante will take an interest in the coding and with his access to all his prodidious data - out will pop a very convincing FM. i like the idea of programmer where somebody else does all the work :-) it is possible I might give some further thought to terrain improvements.

took some time this, hours and hours. I'm looking forward to lobbing a sidewinder at it. I think we'll first complete more of the radar. its got some text stuff missing and a horizon cue - things like that.

now we try to model the bluefox STT mode. still stuff to be done, the elevation level is missing and a bunch of other boring code things to make it work more properly but at least it looks right? its amazing how much progress we've made this week, mainly because i've been off work sick with mumps.

yesterdays multiple blips were script coded today I advanced things a little and got them to track 'real' (but stationary) objects www.thunder-works.com/media/radar.avi This is a good test because now we have confidence that our system will display any number of contacts. in general with regards to the game as a whole - I realise that script programming is a wonderful thing.

today got the radar to plot more than one blip. this is all very exciting, once we've modelled one advanced cockpit like the harrier without problems, then we can take this and apply to the others. never got so excited over three virtual radar blips before.

website traffic was very high :-) supprisingly high http://www.avsim.com/ this site put a link to us and the surge in visitors killed us. about 1 hit every two seconds. amazing stuff. we had thought that 20gig limit would normally be sufficient.

never the less. we think that they are not invincible aircraft. difficult and trickski for sure though. there is a bit of a feeling that the argies could do better. sim pilots do a lot of virtual flying and might be sort of ace like in comparison to the guys who were flung against the brits. They had some very well trained units, but they were held in reserve to defend the capital and to deter chilean agression. From what we've gathered, the pilots who went upgainst the sea harrier pilots were not intensively trained in ACM whereas sea harrier pilots, there main task was supposed to be fleet protection - interception. in war games, french mirages made mincemeat out of harriers.

graphics. yes these look quite good and will get better. There are some important lacking things still yet to finish hopefull in next couple of months - things like terrain features. I think we've got grass done very well - but we need also to be able to do roads and buildings and things like that. edit oh yeah, we've got experimental roads though - tile based approach, but it looks like of like.. ..a tile based a approach. a bit falconish - not the sort of look we'd been wanting, instead we will be doing draping roads in some way.

working on radar now.

sort this out later - its something I should have done actually as I dont think you've fully figured out the LUA scripting. its just a bit of dial script code that needs changing, its nothing. but yeah, thanks for pointing it out - i'd not have seen it for ages otherwise.

actually there is a sort of online issue that might slow down play. obviously i am speculating, but it might be like our skyhawk is internally - basicall a harrier the nossle command paths disconnected. (and a helecoptor will also be a harrier with nossle commands disconnected :) ) - well no doubt there would be some sort of net code being passed about. normally in a wwii type game you cant see planes beyond a certain distanced so updates never need to be given - perhaps in the lomac net code information on all aircraft is being updated to ensure that the radar simulation can pick out what it can see. so in a battle between SU25's and A10's the internal model might be just the same still - its a craft with radar displays switched off, but it may have been simpler for the programmers to have the same routines bobbing about. of course, I'm just guesses - its what I'd be tempted to do, and what I'd definately do if time was short. it is true, I do prefer online play than single play, however - I very much enjoyed IL's series of missions, played once, you can pretend they're a campaign, its just boring to do it ever again. Infact, I found it very atmospheric and I thought philosophical thoughts as I flew my 109 against the russians listening to the excited and seemingly well acted radio chat... ...the sim with a true campaign I've played is Birds Of Prey (Amiga) - as simple as it was, i found it engrosing. So I can well understand the desire for a campaign in jet thunder. And so there will be a dynamic campaign in it. I like a challenge, and online play is sort of, make your own fun - its code wise - apart from the network code - kind of simpler, less challenging. I am really looking forward to when we start working more thoughourly on the campaign system. We may expand the team then. we are using a lot of scripting now, more and more and game play script programmers might be what we need. as for the subject of Strike Fighters project 1. which I also have. Its kind of nice, I've some fun combat in it. Its main problem, is in my opinion - the difficulting in seeing other aircraft, i think they aircraft may be depth culled to invisibility too soon or something. I think the main reason why I never got into falcon 4 was because I lost the CD - I was going to buy another copy but could never find one. The improvements the comunity made - in particular, the higher resolution mode was to my mind, impressive. But really, I needed a teacher to tell me how to play it.

http://www.thunder-works.com/movies/jumptest2.avi a more recent more complete cockpit

yeah. lol. Kind of wish we were doing a wwi game. so much easier. rotating dials - how hard is that? not very. I agree with you about popularity of wwii simming. I myself never got into lomac or even falcon4. complicated for my little mind. the realistic simplicity of IL2 was far more too my liking.. ..however, in jet thunder - we have a sort of inherent compromise. The avionics are not that sophisticated. in the harrier case, I think some of it is 'broken' undevelopped, like some sources say that sidewinders may have had no radar cue - simple limited to uncaged mode. both navigation and weoponry are very important to get right. we still lack a bit of information so at first the radar will be a generic. dante explains it to me, making sure I do it right: I am thinking that the difference in the sea harriers search mode will be subtle and largely superficial. We will be modelling realistic radar - probably with spherical mapped radar cross sections (and a similar approach to infrared cross section) - maybe it has other do dars and widgets too, like brightness - gain? - if its in, we'll try and model it. I suppose this and the weapons has a sort of priority because we find ourselves doing this now and the nav stuff later - but we want it done right. maybe with an IL2 like GPS for newbies - beacons and things like that for the keen. what I am excited about with this is that its going to have some good potential for online scraps. realistic and yet fun due to what I imagine to be accessibility of relatively simple - not much above - WWII avionics. the Skyhawk which along with the Sea Harrier will feature in a demo in the near future are quite simple in their avionics. The Skyhawk itself doesnt even have a radar. :-) just what appears to be a WWII style gunsight. And with aircombat being up close and personal, it should appeal, I hope to the both fans of jet sims and wwii simmers who are in it for the close combat. the scripting approach we use makes it possible to model a lot of different electronic cockpits with lots of function.

well happy new year people. today I've been working on jet engine model. quite simple at the moment but much better than our previous instant throttle linked thrust model. now with lua scripted, the behaviour of the jet engine is modelled - albeit simply - and yet since it is script it is refinable. perhaps I am doing this more or less right. there is min and max fuelflow rates for engine RPM. if the desired rpm is above or below current rpm, fuel flow rates are modified within the parameters allowed. currently the model is quite simple and behaving linearly the faster the turbine turns, the more drag there is on it - the more fuel flow, the more force accelerating it. Had some fun things, can imagine it happening when people have been developing turbines in the past - were the fuel flow constraints are perhaps not too well set and the engine just keeps on getting more and more power and there is nothing you can do :-) it is therefore more responsive at higher rpm - so its seems to work quite nicely. this lets me set the gauges now.

we continue to improve cockpit, now that I become focused on the the gauges: fuel flow and engine RPM for the harrier I am not quite sure what happens. i presume the engine is started and it ticks over at some RPM or other. Then there is fuel flow and rpm. I presume throttle regulates fuel flow (fairly rapidly) and then engine RPM climbs or falls in some way. - there may also be a delay before a surge in fuel flow gets to affect the engine? until now its just been, instant, throttle sets the thrust in the FM. so if anybody has any thoughts on this, release them here :)

you seem to know your stuff. I dont suppose you can program in C++ or LUA? :-) I am rapidly trying to develop stuff now so will create a system that somewhat resembles how things should be with view of returning to perfect things.

perhaps we will explore that idea. understand that a version of falcon is using the x-plane system and reports of it are very good. main problem is me, I like to basically do everything program wise.

oh yeah head shake, that can be added in at any time really, not much to it. pretty cool that head shake behaviour eh? the shaking shows off the 3d nature of the cockpit.

just done a dev movie. its not yet finished, so its not on our front page - but it is here for those that wonder. clearly, there is some little glitches and incomplete behaviour - but the movie illustrates current stage. we develop things separate. we have a fledgling mission system an example of the engine in a game setting were you can take off and land from hermes, drop bombs and things like that. and this is other segment is where we continue to focus hard on the cockpit. It takes a while for files to load up so when tweaking things it good to separate things out a bit. Here in this movie we fly over a featureless flat grassland. Within this simplified world the physics, and the avionics can be modelled quite well. xvid codec required http://www.thunder-works.com/media/seaharrierinprogress.avi

its getting exciting now. the way i see it all major things we now have a handle on. Previously, there was still epic mile stones to tackle - now its all coasting downhill now - mostly content creation, tweaking and completion. what do you people think of this variation - i changed dantes textures slightly. i was trying to match the colours of a real photo anyway, this is not that important - what we've been doing recently is continueing to add function to the cockpit. more dials doing their thing, more hud details. It is going to be quite an undertaking all this. next we must model different hud modes, and the navigation and radar gadgetry.

stuff still to do. most of the dials can now be programmed by lua script. For the first time (slow I know) we've got a artificial horizon dial. It took so long because for some reason I'd never got round to the relatively simple maths of finding the a,b,c alignment of an object given its matrix. this raises some interesting possibilities to my mind. a,b,c alignment seems to be a jolly reasonable way of tweening between alignments. Will be useful for when we have a track recording function and internet. our previous internet code did have a tweened position but not a tweened alignment. hmm on 2nd thoughts that might lead to some odd effects when the aircraft is pointing upwards very near the vertical axis