Open Source Space Flight Combat sim

Something just tells me this isn't getting off the ground. It seems far too ambitious.

Plus, I can identify with the 'explain in non-programmer terms' ... sure, I get it, but Kazan's not too hot on explaining in anything else, apparently.

Still, I'll withhold judgment until I see some more stuff, it wouldn't be fair otherwise.
 
star: if i felt like taking longer to use non-programmer jargon i could

"cononical features" is programmer jargon for the minimum features required to function


i have the cononical features of opship and opmodel working, im working on a freespace 2 POF to opship/opmodel couplet converter so i can have some test data - my physics and collision detection programers are theorizing - i have three programmers without assignments ATM, im going over to one of them's place tonight. After I convey my plans for the engine to her i'll bring her up to speed with game programming and give her an assignment.

I am planning on holding a general meeting of my programmers next week to official kick off game developement
 
ok.. i've put a prettier skin on the Game Type Selection Screen [mod selection]

http://www.public.iastate.edu/~kazan/OP00.png
http://www.public.iastate.edu/~kazan/OP01.png
http://www.public.iastate.edu/~kazan/OP02.png
http://www.public.iastate.edu/~kazan/OP03.png

this is by no means the final one - but this is one of the things int he game that individual mods cannot affect - only the copy in data/root_op/... will be used

i am currently working on a FreeSpace 2 POF to opmodel converter [which will be extended into an opmodel/opship couplet converter from the same format]

i just have to write the routines to write the model to disk.. then i can start horsing around with the renderer and stuff
 
stuff

this is something i sent to the contacts email address, but never got a reply.

as sent :

"

this is just a set of comments/opinions i have. just for another view.

make open privateer multiplayer only, with a 'fake-me-out single player'
- single player is a local client connected to local server, with a local state set dedicated to 'pseudo single player'
- this allows for simultaneous multiplayer and single player development in a single effort.
- you dont have to go back and modify what you've done to add-in multiplayer
- multiplayer will natively support missions.
- allow people to join your single player game and play with you as a hired wingman.
- possibly have a cooperative mode where there are 2 protaginists, or N protaginists in a multiplayer mode, seeking to complete the main mission first.

muliplayer section
- have a CS - P2P hybrid.
here is a sample scenario (same as i mentioned to hellcatv) :

"

the benefit of P2P broadcasts are 1/2 the lag time from the moment a person fires, to the time a shot is seen.

this is a clearer overview :
(caps A/B are entitys)

A sends coordinate/velocity/accel to server
B sends coordinate/velocity/accel to server
Server returns all coordinates to all other players withing sensor ranges of relative other players.

A broadcasts coordinate/velocity/accel to clients in sensor range
B broadcasts coordinate/velocity/accel to clients in sensor range
NOTE : this might be done at a range closer than sensor range, i.e. for faster and smoother gameplay, clients can update directly if REALLY close, like within 1/2 gun range or soimething.

A shoots at B.
- broadcast fire to all within sensor range

A strikes B on client side
- broadcast strike to all within sensor range
- send strike to server.
* server evaluates B's position at the time A sent the strike, back tracking B's position X seconds (where X = latency between A and B)
* server sends a validated health change to B OR server nullifies message of A, sending an inconsistency message. (accumulation of messages will result in a drop - aka kick cheater after he keeps it up)

N minute time interval.
- global server updated on player healths, states, etc. their active accounts stored.

A lands on planet, saves.
- global server stores a restore point.
- players allowed some amount of restores before they must continue as-is (or be dead and restart if thats what is). maybe have restore credits count up every so often, like one-more a month.

the broadcasting wouldnt be a problem if its only to ships in sensor range. this range could be cut down some. maybe have action-sensors and navigation-sensors be a bit off. like you can lock planets, etc, from far off, but you wont detect an enemy ship till some bit closer. in any case, it would smoothe out gameplay, while still having a server maintain fairness.

the only thing with a CS only setup is that you are dependant on your connection to that server completely. if you have bad ping to the server, you're screwed. while you might have good ping to many of the other players. [in a hybrid system] Even if you had terrible ping to the server, and average to the players, your gameplay would still be average, and your shots would take off health at some delay after they strike. but that wouldnt be that terrible in exchange for smoothe play.

-scheherazade

"

this thread has the whole post http://sourceforge.net/forum/forum.php?thread_id=750923&forum_id=205254

i know of the P2P bandwidth issue. though if the P2P connections were capped to the immediate area, and the rest handled CS style, there would be a benefit of lesser lag (p2p) where it will be seen, and global containment like in CS. i think that rather than choose one ideology, take the best out of both. a new approach might do well. no one has tried, so it might be good, never know.



also, on the modellig end, give per pixel lighting a go (like tenebrae http://tenebrae.sourceforge.net/index.php?page=screenshots.txt)
and possibly scripted shader support for modellers. (like quake3)
supoprt 32 bit tga's

these options would :
- look f'ing sweet and attract peoples attention like gnats to fly-tape.
- give artists some freedom to work with more ideas.
- have included alpha channel for transparency in textures.
- bump maps for surface details, saving polys.
- gloss maps for reflective ares, like glass, or some evvects on rounded edges to exaggerate backdrop lighting, like a suns rays creeping around edges of a ship from behind.


if i were to suggest a model format, i would say md3 (like in quake3)
why? because it is a 3ds file (arguably the most common format) with long filename support and extended frame cap added in. (or so i've read). i like 3ds files cause they are super universal and fairly versetile. the long filenames would be nice, and the frame cap is pretty low on them. also md3 documentation is fairly ez to get.

also, like in quake3 (btw i have worked on and still do work on quake mods, so i can attest to some advantages) i would recommend the use of tags. (just model objects named tag_blah, with a certain dimension that can be used as a new origin for a model daisy-chained to the tag).
this way you can :
- have visible weapons support. guns/missiles attached to tag locations. buy/sell/upgrade arms would be visible.
- have automatically distributed guns and calculated available gun slots and expansion slots
- have missiles fire from tag coordinates, saving effort of having to code in the coords or script them
- cut down on animations if any. ex. an attached model for landing gear would have its own anims. if the body had anims for some thing like a weapons bay opening, there would be separate anim sets.
i.e. body frames 0-15, gear frames 0-15
rather than
total frames 0-64, where 1-15 = gear up bay closed, 16-31 = gear down bay closed, 32-47 = gear up bay open, 48-64 = gear down bay open.
tags and attached models make animation simpler, modelling time shorter, and file size smaller.

so anyways, thats my rant. what are your feelings.

-scheherazade

"
 
So, is there any chance of turning this into the massive multiplayer online space simulation that Freelancer was originally meant to be?
 
kazan <

ya do SP first.

i mean archietecture wise, if you make the game a local-local server, and a local client connecting to his own machine, you'd be getting 2 birds with 1 stone.

you start with separated physics and UI. which is a pain to separate later for MP, and usually results in bad MP playability cause its not really optimized for it.

you're still working on single player. but then to get multiplayer working its just a little bit of work and you're there. while otherwise you're recoding a massive amount.

-schehearzade
 
btw, if you want to do realtime lighting with bumpmaps gloss maps, etc. and support Vwep etc, i'm prepared to make and modify models for you to be format compatible with the engine.

also i've done a bit of this sorta deal for flight sim mods. so i have a pretty good idea of how to go about making models modular and functional.

would be great to have landing sequences like coming down, pulling out the gear, setting down. rather than the usual disappear and reappear in some location.

also, if you set up the lighting engine good, you can do this :

when landed*
- each location has a set of camera locations.
- when you enter an area in the location, your screen is a viewport of a certain camera.
- over this viewport is drawn a 3d model of the place. it is lit with the ingame engine.
- windows in the model let you see through it, seeing the viewport.

this would result in something like :
- you are in perry naval, and you are in the launch bay
-- you see the actual space outside spinning
- you are in the concourse at perry
-- you see whats really flying by in the station windows.

-scheherazade
 
that's an interesting idea... i may have to take a look at doing a little bit of that [the perry navel idea]

remember - my engine isn't OFFICIALLY wing commander... im just giving very big preference to modability and i'd love to see a Wing Commander mod for it [i would provide tons of support]

i started a mapping contest for making a map for OpenPrivateer's yet unnamed "Official" campaign/universe... my entry to the contest [partially completed]

[lvlshot]http://www.public.iastate.edu/~kazan/StarMapProgress.jpg[/lvlshot]
 
looks good.

are you making the engine with the intention of there being a WC implementation of it?

if so you might as well start with a WC implementation.

or make it and the wc mod in parallel. so that both are done at the same time.

btw, i made a drayman model you might use. i am gonna clean it up later. i'm doing all the wcp civilian ships. when i'm done with them i'll send them to you.

-scheherazade
 
i'll be giving the WC Priv mod a lot of logistic support.. but i am making an open source engine, and the game story patents, etc are going to belong to me and my team because we do want some $$$$

but yes i do have a WC implementation in mind for much fun :D
 
he's talking about making $ on the engine as implemented originally, not off a possible wc version.

oh ya, if you make all your art from scratch, and dont use identical names, you dont enfringe copyright.

-scheherazade
 
the only names i've used the same that could be even claimed any connection is Akwende Drive... welll... im afraid far too many people use that in non-WC related terms for them to do shit

the engine is open source.. but the game will be making us $$$ hopefully
 
Back
Top