scary_pigeon

continueing from a discussion with dante the other day - he requested some extra terrain effects. Dante should be able to enhance the implementation later with different textures and settings and such. but the effect looks quite good. Unplanned is the apparant shallow water effect.

i wonder if there is any reason for us even to write to the z buffer for rendering sea?

realised that lightsourcing was missing from terrain. makes a difference sure enough. i think this looks much better.

for the corrious - it was loading all terrain because currently our new terrain is only about 80km and the view distance happened to be 80km - now with the view distance reduced we have as expected lots of wonderful load on demand pop up. :-) what is nice is that is not going to be hard to include islands in the middle of atlantic. which makes vulcan flights possible rather than air spawning.

thats what page files are for. :P

i doubt it is doing exactly what you think it is doing. what it is supposed to do - is loads up the positions of all terrain elements in to magic particlemesh data structure. Then comes the render phase. it selects a region, checks if it has been loaded yet - if has displays it, if not loads it. i will check it after though to make sure its not doing anything perculiar.

i dissabled the tweening for testing purposes. glad to see people still interested. the last few months have been slow. network code took 3 months rather than 3 weeks.

weather effects are not any more developed than what we saw in previous screen shots - sort of fluffy clouds which can be darkened and interesting skydome effects along with rain. fog and haze type things are easy enough to do.

definately, it adds to the sort of authenticity in a small way perhaps - like how in IL2 the voices are all spoken in german and russian rather than accented english.

another missile test - this time we are working on a new terrain: http://thunder-works.com/media/terrainwip.avi it will improve yet. this is what we did last night. later on some more work on damage actually killing an aircraft. the other day i did manage to with a missile knock off bits of a sea harrier (by mistake - sort of firing the missile blindly into a an ai dogfight) - but it carried on flying with the back half of hte plane missing. so that wont do.

super. so, i read that this project is aligning itself with that new falcon game. is that to be an exclusive arangement? some of what you've got would be very very useful. i wonder how we would use the mesh files? probably we'd have to get one of your coders to help me on that as we've sort of ended up using an odd approach that grew out of not knowing what were doing from years gone back. we use multple .asc files for vertex information - a msh file for u,v information and script files to link objects together - oh and axis of rotations are defined as last triangle as axis :-) so its all a bit strange.

most recently i have crushed an error.. the missile . thaat didnt excplode completelu- it what do that. boom. bugt carry on flying no longer though in latest code.

another movie xvid codec needed http://fluffysteve.pwp.blueyonder.co.uk/air2airb.avi what we see here is a test of the missile exploding after tracking some target. The way the project has ended up is that everything is played over an net connection weather across the world or to a local server. The skyhawk the missile that was fired at was a bot aircraft. the missile too was a bot flying under the servers direction in its single minded pursuit. missions are defined on the server as will campaigns. much remains, in the movie you can see how the missile carries on through the target - this is because i've not programmed it to be removed properly from the client: it just takes its last velocity and accel information and then flies. Also, its just fireworks at the moment as the missile does no damage yet. These are not hard things to do and will be done shortly. probably when i get back from work :)

magnificient. i think the only flaw is low level look - the swimmy blury textures.

this i think is a first for us in some ways - for some months we've had bots firing missiles, then it was the player able to fire dumb fire missiles. Now I did a test firing. it actually missed the plane in that attempt.

for the sake of properganda mainly - we worked today on graphical stuff. still in the test shell though: http://thunder-works.com/images/missiletests5.avi this is a movie of a missile hitting some object on the ground. What we were seeking to do was to put some meaningful placeholder effects for missile trails and explosions. the game engine issue that is most prominant now is solving a client crash when missile are fired over the internet. When that is done we'll be putting back in the landscape - most likely a new approach.

hehe - right now we have a interface which is an msdos prompt window and when i tested guided missiles sometimes the server would crash :) but i think its really the right thing to do to prioritise netcode before the last minute!

today i've been enjoying my new monitor - a 19" one with higher resolution than previous - its pleasant as when coding you can see much more source - less time consuming scrolling of text occurs. and as for flying - its suddenly a lot more fun. my favourate game at the moment remains the imperfect but playable wwii online. it is nice to be able to see more easily infantry scurrying around the ground. so now we have guided missiles firable from human players - unlike bullets, they currently do no damage. i think the way to program this is to go for a bit of conformity in the code and say there are to be such thing as High Explosive events as just another message to pass round. this should mean that the same sort of thing can apply to bombs or whatever.

indeed, so far in such discussion dante has continued to be right -3 years or so ago i was saying, "ahh, only another 6 months till completion." I tend to be highly optimistic about such things.

2 more years till gold release? i dont think it will take anything like that long. when the demo is done, i see mainly only content creation to follow and much of that could come in free patches like IL2. quite a bit more is there that is not active yet in the net code -for example spawn points - these are lua scripted in the server mission script - these are then downloaded and any of these can be selected.. ..its just we're currently running in an interfaceless test shell so hidden from dante's view is the download of the spawn points, the spawn selection, the unique arming process, etc. next important thing now is to finish of the network missile thing so not just AI can fire missiles, then add terrain - and then we've got a demo basically - in particular, when the mirage cockpit is finished and the luascripts for that cockpit programmed.

well i found that screenshot of the old game kind of inspirational. the site of the warships on the move, looks simple but purposeful. project wise, still stuck on the net code. the particular bug is a message related thing were sometimes chat messages/game events get lost - some server bug. my most recent attempt to fix this was to simply the code - the net code i created was in some ways i think far too adventurous - with aspects of future projects in mind like World War III online- massively multiplayer war game stuff. the more complicated something is - the more there is to go wrong - so now in debugging i bipass some of the clever but potential faulty stuff and fall back on simpler methods: the bug appears to possible occur around some bit of code relating to the decisions on what messages to pass on to which clients for example: one client holds the trigger down - there server learns of this. then must decide who, out potentially hundrends of clients need know about that event. this cuts down on bandwidth when there are many players spread over a large virtual terrain; i've fallen back on simply sending everybody the message. it seems to work, but so it did seem to work before - a sporadic bug that i've spent weeks sulkingly trying to pin down - to point of a sort of dispair where I give for a day and have some fun: i built a mount for a joystick to stop it wobbling over the table. I enjoyed the physical effort of cutting through the thick planks. i've also been making balsa airplanes telling myself that it is sort of project related :-)

my favourate flightsim remains IL2. its graphics and flight model I found amazing when it was first released. it is my main inspiration.

slowly but surely the demo approaches soon. hit a bit of a rock recently code bugs. i think though in the next few days, our small dev team will be testing shooting each other down. there is still much to do. yrrd who has joined us, and has been tasked with perfecting the fm, has so far not perfected it. And then there are millions of other little things to do. I think though that what is left is millions of little things rather than big huge question mark, 'how on earth r we going to do that?' sort of items. the amusing thing today was in frustration at the bugs i held out my code psychic hand and scrolled the text waiting for my hand to feel the bug beneith and i found in some maths relating to how much bytes to expect to receive in the net code - a missing set of bracket. i find that in the later parts of this projects i have abandoned planning of code and being going on instinct. the main frustration though is that we all in this project must work for a living and not be yet professional 100% on it. distraction is an easy hazard for us I think - and I also expect that it is a similar problem that aflicts and slows down other projects. I just hope that by the time we release - we're not too dated and that it meets expectations.

collision detection has been around in our code for years. but recently we've beenmaking a lot of changes to the code - a sort of step back do things a bit better. this is running over a network where I pick on an skyhawk on the ground. http://fluffysteve.pwp.blueyonder.co.uk/testforweb.avi

cheers. its my birthday next thursday - and it creates a sort of less than arbitary date - i'll be 27 soon be 30 :) (dante never seems to update my age on the website - im forever 25 hehe) anyway this creates a date were i want to have some significant progress towards something more playable. rightnow we can shoot each other down - but because of incompleteness, we'll be flying around in invisible planes or with bits missing like wings and stuff with no adverse effects on the flight behaviour - of course - this cant carry on! today i wrote more netcode relating to position tweening. we had a half done tween process based on dead reckoning - but the positions you project from need to be tweened otherwise you get a teleport and glide effect - whereas we want a constant smooth but hopefully accurate transition.