Contributing to VegaStrike
-
- Insys Pilot
- Posts: 2
- Joined: Sat Nov 25, 2006 12:52 pm
- Contact:
Contributing to VegaStrike
Hello,
I would like to release my creativity (mainly) in the C++ parts of VegaStrike. I'm particularly interested in (no special order):
* graphical representation of damage on HUD.
* more realistic damage model.
* making the relationship of scalings more realistic (the sun in the starting system is not too much bigger than my ship)
* enhancing performance of graphics engine (visible ship parts lag behind crosshair)
* better font rendering.
* amending joystick support (choose joystick, map keys from a GUI, option to make axes act logarithmically) and configuration in general (pretty ad-hoc XML/Custom format right now).
* enhancing the Linux port (I miss a working vssetup and vslauncher in the latest release).
* co-operative multiplayer.
* kind-of-realistic landing on planets/in docks
* making the point-and-click parts more atmospheric, better to read and more beautiful (it's hard to read a conversation when there's only one line of text at a time). That might be a Python or C++/Python job as far as I have seen.
Given some time, I can surely think of many more things to add or modify (I'm sure we all can ).
Please say so if one of those is already in current SVN -- all this is based on the latest release -- 0.4.3 IIRC.
I also wonder about frameworkish things:
* why is the SVN repository so cluttered? I see a lot of side projects (evilwm, ethereal, ...) in current trunk/, and there are huge binary files in in. That's pretty bad practice and, what's more grave, isn't going to help attracting contributors :/
* why are the mailing lists dead (well, if you don't count the spam)? They are a very efficient means of communication in contrast to a BBS, and especially developers are fond of them.
* why was it decided to switch to OGRE? There's at least on other fine engine, Crystal Space, that you should be aware of. However, I can't seem to find any discussion on the advantages and disadvantages of them.
If you need someone to manage things like that, I might also be able to help -- I have been maintainer of an established project (GNU Parted) for the past ~18 months and am pretty through all project management issues).
Thank you.
Rylis
I would like to release my creativity (mainly) in the C++ parts of VegaStrike. I'm particularly interested in (no special order):
* graphical representation of damage on HUD.
* more realistic damage model.
* making the relationship of scalings more realistic (the sun in the starting system is not too much bigger than my ship)
* enhancing performance of graphics engine (visible ship parts lag behind crosshair)
* better font rendering.
* amending joystick support (choose joystick, map keys from a GUI, option to make axes act logarithmically) and configuration in general (pretty ad-hoc XML/Custom format right now).
* enhancing the Linux port (I miss a working vssetup and vslauncher in the latest release).
* co-operative multiplayer.
* kind-of-realistic landing on planets/in docks
* making the point-and-click parts more atmospheric, better to read and more beautiful (it's hard to read a conversation when there's only one line of text at a time). That might be a Python or C++/Python job as far as I have seen.
Given some time, I can surely think of many more things to add or modify (I'm sure we all can ).
Please say so if one of those is already in current SVN -- all this is based on the latest release -- 0.4.3 IIRC.
I also wonder about frameworkish things:
* why is the SVN repository so cluttered? I see a lot of side projects (evilwm, ethereal, ...) in current trunk/, and there are huge binary files in in. That's pretty bad practice and, what's more grave, isn't going to help attracting contributors :/
* why are the mailing lists dead (well, if you don't count the spam)? They are a very efficient means of communication in contrast to a BBS, and especially developers are fond of them.
* why was it decided to switch to OGRE? There's at least on other fine engine, Crystal Space, that you should be aware of. However, I can't seem to find any discussion on the advantages and disadvantages of them.
If you need someone to manage things like that, I might also be able to help -- I have been maintainer of an established project (GNU Parted) for the past ~18 months and am pretty through all project management issues).
Thank you.
Rylis
-
- Developer
- Posts: 3980
- Joined: Fri Jan 03, 2003 4:53 am
- Location: Stanford, CA
- Contact:
Hi Rylis:
it sounds like you have some very good ideas here and we could certainly use another coder, especially at a time like this when folks are busy at school
Ogre will take care of 90% of the graphical changes you suggest here--so one idea would be for you to help out with ogre---because it can really use another dedicated developer working on it---and ogre has builtin facilities for most of these things (i.e. better fonts, more performant graphics engine, etc)
As for why we chose ogre... well I originally planned to use crystal space, but the project wasn't nearly as active in keeping up with all the evolution of graphics API's.... ogre had a working system for stencil shadows and pixel shaders with crystal space only promising it in the distant future.... now I think crystal space delivered by now--- but I think ogre keeps taking baby steps towards newer graphical features whereas crystal space has to worry about physics and other things we'd end up ignoring (space physics is a lot different than land physics--esp wrt double precision, etc)
let me treat each of your individual points:
a) damage on the hud: that would be quite useful--we'd probably consider a better internal representation of the damage--though SVN does have a mode where it lists damaged systems---the method for figuring these out is quite poor
b) see a
c) I recall that most systems in VS are actually quite accurate---at least Sol is 1/10 normal size---which means the sun is huge.... which system has the scale problem? that's one thing I think we've dealt with quite effectively--and we're working on even larger systems for the next release...
d) the visible lag I think is an effect that can be disabled in the config file (hud_lag?)
e) better fonts wouldn't be terribly hard---but given that ogre already has these--may as well just work on ogre?
f) better joystick support would be awesome--and probably a reasonable first project to familiarize yourself with things...
g) hmm vssetup worked for me--but ya a few folks had it crash on load---I think it had something to do with the -static flag I passed to gcc
Now what I think a great starter project could be is a new cross platform config file editor that lets folks set the XML to various presets, but also lets folks customize individual values in the XML aside from the various presets (so they can fiddle with things individually after selecting a decent preset)
h) yes multiplayer development would be awesome--right now it's just deathmatch--but with some cargo and upgrades it could be a lot more like privateer--and with some sort of mutual goals it could become cooperative---course it would only be a few lines of code to make a coop multiplayer gauntlet that spawns enemy bots one by one with increasing difficulty---spiritplumber's WC server has bots flying around that respawn ---so using that code would be easy
i) realistic landing---probably needs a huge graphics overhaul so should wait until ogre work is ready
j) definitely worth doing at some point---not sure it's a huge tangible gameplay benefit--but it could make folks more interested in seeing the rest of the game
ahh yes--- 0.4.3. is sorely out of date--and it's strongly recommended you try out the svn release---or at least some of the mods to the games' releases--because they have newer engines I think even...
frameworks:
clutter? side projects should be removed--but they were convenient to have there at the time... not sure why large binary files are there---- we can look at that...
as for mailing lists---it's dead because of the spam mostly :-/ we turned it to member only postings---- but somehow the spam still gets through... we can definitely coordinate there--and it's a good place to post ideas---we can revive it from the dead if there's interest!
I addressed the ogre thing earlier
definitely would *welcome* any administrative help on any of this stuff--- a lot of things are the way they are because of inertia---because when the developers have an hour of time they'd rather spend it on working the engine rather than cleaning up management stuff that hobbles along well enough as is...not a great policy--but we do what we individually like cus it's a volunteer project (for worse and better)....
hopefully this answers some of your questions--- the first two steps I'd suggest are to check out both the main trunk and the ogre forks from svn---
svn co https://svn.sourceforge.net/svnroot/veg ... vegastrike
svn co https://svn.sourceforge.net/svnroot/veg ... nk/data4.x
svn co https://svn.sourceforge.net/svnroot/veg ... nk/vssetup
svn co https://svn.sourceforge.net/svnroot/veg ... gre_branch
and try them out individually---decide which side of the fork you want to work on (i.e. multiplayer, etc on the active side--or graphics etc on the ogre side) and either post back here or fire up a mailing list post and we'll see if we can revive that
it sounds like you have some very good ideas here and we could certainly use another coder, especially at a time like this when folks are busy at school
Ogre will take care of 90% of the graphical changes you suggest here--so one idea would be for you to help out with ogre---because it can really use another dedicated developer working on it---and ogre has builtin facilities for most of these things (i.e. better fonts, more performant graphics engine, etc)
As for why we chose ogre... well I originally planned to use crystal space, but the project wasn't nearly as active in keeping up with all the evolution of graphics API's.... ogre had a working system for stencil shadows and pixel shaders with crystal space only promising it in the distant future.... now I think crystal space delivered by now--- but I think ogre keeps taking baby steps towards newer graphical features whereas crystal space has to worry about physics and other things we'd end up ignoring (space physics is a lot different than land physics--esp wrt double precision, etc)
let me treat each of your individual points:
a) damage on the hud: that would be quite useful--we'd probably consider a better internal representation of the damage--though SVN does have a mode where it lists damaged systems---the method for figuring these out is quite poor
b) see a
c) I recall that most systems in VS are actually quite accurate---at least Sol is 1/10 normal size---which means the sun is huge.... which system has the scale problem? that's one thing I think we've dealt with quite effectively--and we're working on even larger systems for the next release...
d) the visible lag I think is an effect that can be disabled in the config file (hud_lag?)
e) better fonts wouldn't be terribly hard---but given that ogre already has these--may as well just work on ogre?
f) better joystick support would be awesome--and probably a reasonable first project to familiarize yourself with things...
g) hmm vssetup worked for me--but ya a few folks had it crash on load---I think it had something to do with the -static flag I passed to gcc
Now what I think a great starter project could be is a new cross platform config file editor that lets folks set the XML to various presets, but also lets folks customize individual values in the XML aside from the various presets (so they can fiddle with things individually after selecting a decent preset)
h) yes multiplayer development would be awesome--right now it's just deathmatch--but with some cargo and upgrades it could be a lot more like privateer--and with some sort of mutual goals it could become cooperative---course it would only be a few lines of code to make a coop multiplayer gauntlet that spawns enemy bots one by one with increasing difficulty---spiritplumber's WC server has bots flying around that respawn ---so using that code would be easy
i) realistic landing---probably needs a huge graphics overhaul so should wait until ogre work is ready
j) definitely worth doing at some point---not sure it's a huge tangible gameplay benefit--but it could make folks more interested in seeing the rest of the game
ahh yes--- 0.4.3. is sorely out of date--and it's strongly recommended you try out the svn release---or at least some of the mods to the games' releases--because they have newer engines I think even...
frameworks:
clutter? side projects should be removed--but they were convenient to have there at the time... not sure why large binary files are there---- we can look at that...
as for mailing lists---it's dead because of the spam mostly :-/ we turned it to member only postings---- but somehow the spam still gets through... we can definitely coordinate there--and it's a good place to post ideas---we can revive it from the dead if there's interest!
I addressed the ogre thing earlier
definitely would *welcome* any administrative help on any of this stuff--- a lot of things are the way they are because of inertia---because when the developers have an hour of time they'd rather spend it on working the engine rather than cleaning up management stuff that hobbles along well enough as is...not a great policy--but we do what we individually like cus it's a volunteer project (for worse and better)....
hopefully this answers some of your questions--- the first two steps I'd suggest are to check out both the main trunk and the ogre forks from svn---
svn co https://svn.sourceforge.net/svnroot/veg ... vegastrike
svn co https://svn.sourceforge.net/svnroot/veg ... nk/data4.x
svn co https://svn.sourceforge.net/svnroot/veg ... nk/vssetup
svn co https://svn.sourceforge.net/svnroot/veg ... gre_branch
and try them out individually---decide which side of the fork you want to work on (i.e. multiplayer, etc on the active side--or graphics etc on the ogre side) and either post back here or fire up a mailing list post and we'll see if we can revive that
Vega Strike Lead Developer
http://vegastrike.sourceforge.net/
http://vegastrike.sourceforge.net/
-
- Elite
- Posts: 7243
- Joined: Mon Apr 18, 2005 2:40 pm
- Location: LS87, Buenos Aires, República Argentina
Re: Contributing to VegaStrike
All graphical tasks would be better deferred to the Ogre port. IE: Concentrate on switching the graphical engine. Why? Because the graphical engine needed a rewrite already, and effort is better spent making the new framework handle whatever feature you want, rather than adding the feature to the current engine. That is, unless the feature is easily added, and is worth the effort (ie - is good for an imminent release).Rylis Blueshift wrote: * graphical representation of damage on HUD.
* enhancing performance of graphics engine (visible ship parts lag behind crosshair)
* better font rendering.
Lots of ideas on this front, but little time to code them. I'd be happy to discuss them, over mail probably (and be patient, I'm very busy lately, and so it may be a slow discussion).Rylis Blueshift wrote: * more realistic damage model.
You lost me. The sun is really bigger than your ship. Really. Even in 0.4.3.Rylis Blueshift wrote: * making the relationship of scalings more realistic (the sun in the starting system is not too much bigger than my ship)
There's work on deadmatch multiplayer. That kind of multiplayer is the first step, it allows to debug and tweak the underlying system. From that, co-operative is just a gameplay twist.Rylis Blueshift wrote: * co-operative multiplayer.
Would love that. Also, some ideas exist on this front, but no time to code them. Mostly, because they need a lot of AI programming, believe it or not.Rylis Blueshift wrote: * kind-of-realistic landing on planets/in docks
Again, discussion by mail might be the best option.
The binary files are there for two reasons.Rylis Blueshift wrote: * why is the SVN repository so cluttered? I see a lot of side projects (evilwm, ethereal, ...) in current trunk/, and there are huge binary files in in. That's pretty bad practice and, what's more grave, isn't going to help attracting contributors :/
A) They're the game's data files, so they're an important part of the project.
B) If you mean the executables, they provide an easy way for Windows folks to get playable SVN snapshots. Most people on Windows can't build from source. You can configure your SVN sandbox to ignore data4.x/bin if you wish, only windows binaries there.
Nobody has anything to say there. Most of what has to be said is said on the boards... just the way it is. Feel free to use them (the mailing lists), what you say there will be heared just the same.Rylis Blueshift wrote: * why are the mailing lists dead (well, if you don't count the spam)? They are a very efficient means of communication in contrast to a BBS, and especially developers are fond of them.
A) WorksRylis Blueshift wrote: * why was it decided to switch to OGRE? There's at least on other fine engine, Crystal Space, that you should be aware of. However, I can't seem to find any discussion on the advantages and disadvantages of them.
B) Very active development and support
C) Supports the three main platforms VS aims for: Windows, Linux and MacOS.
If Crystal also works, has active development and supports all those platforms... well... there's really no reason, only that one had to be chosen.
-
- Minister of Information
- Posts: 1895
- Joined: Fri Jan 31, 2003 9:40 pm
- Location: The land of tenure (and diaper changes)
Re: Contributing to VegaStrike
Welcome!
I'll try not to dwell too long on points which have adequately been covered by Hellcat and Klauss.
What could use some decided revisiting is the algorithm by which the system generator decides which items to place where in a generated solar system. Admittedly, this is complicated by there not yet being any particular consensus on how many dimensions need to be applied to planet generation, nor how many values in each dimension, in order to achieve the desired spread of object types - in large part because there's still much handwaving as to the desired spread of object types
As for the binaries... some of that is just data set, and there's little to be done about it. Whether or not they continue to abide in their current location, a set of windows executables is a near-necessity, as it allows artisans who are not capable of compiling their own binaries (usually on Windows platforms) to produce content that can be tested on their own end before being submitted for potential inclusion into a data set.
I'll try not to dwell too long on points which have adequately been covered by Hellcat and Klauss.
I'll assume the particulars of that statement are hyperbole, as I know the point itself to be quite valid for the 0.4.3 release, though the statement itself to be inaccurate. System scaling is largely a function of parameters in a given data set, and should require minimal, if any, underlying code changes (I can't remember if there were some orthogonality issues between some of the scaling parameters and pre-built systems - I'll have to check). If you examine the starting system in SVN, you'll notice that you start a comfortable 9 light-minutes or so away from a reasonably sized star. VS happily supports full scaled systems in terms of physical dimensions. Full scale for planetary orbital velocities is a less desireable affair at present, as much for game-play reasons as for anysimulation time granularity issues involving fast moving objects. See also: our model of Sol.Rylis Blueshift wrote: * making the relationship of scalings more realistic (the sun in the starting system is not too much bigger than my ship)
What could use some decided revisiting is the algorithm by which the system generator decides which items to place where in a generated solar system. Admittedly, this is complicated by there not yet being any particular consensus on how many dimensions need to be applied to planet generation, nor how many values in each dimension, in order to achieve the desired spread of object types - in large part because there's still much handwaving as to the desired spread of object types
One man's feature, another man's folly? (At least it's configurable).Rylis Blueshift wrote: * enhancing performance of graphics engine (visible ship parts lag behind crosshair)
Any and all organizational assistance will probably be welcome. Current evidence suggests that the current set of developers do not often find organizational and managerial tasks to be amenably aligned with their forte, interests, or scheduling priorities. Scheduling currently seems to be one of the harder tasks, as we all have fairly imposing outside commitments. For example, of the three current VS sourceforge project admins (1 duplicate account, and one retired) two are PhD students and one is just finishing up secondary school.Rylis Blueshift wrote: I also wonder about frameworkish things:
<snip>
If you need someone to manage things like that, I might also be able to help -- I have been maintainer of an established project (GNU Parted) for the past ~18 months and am pretty through all project management issues).
Good questions. Thankfully, SVN >> CVS, and it wasn't very hard to clean up some of the more egregiously out-of-place cruft (evilwm,ethereal,...). Should be a little tidier now.Rylis Blueshift wrote: * why is the SVN repository so cluttered? I see a lot of side projects (evilwm, ethereal, ...) in current trunk/, and there are huge binary files in in. That's pretty bad practice and, what's more grave, isn't going to help attracting contributors :/
As for the binaries... some of that is just data set, and there's little to be done about it. Whether or not they continue to abide in their current location, a set of windows executables is a near-necessity, as it allows artisans who are not capable of compiling their own binaries (usually on Windows platforms) to produce content that can be tested on their own end before being submitted for potential inclusion into a data set.
Partly, I think it was the spam. Spam saturation depresses interest in using a communications channel. It's a lot cleaner now, so it shouldn't be too hard to rekindle interest. As the mailing lists were unfriendly, given the number of people actively working on any specific part of the project at a given time tending to be a very small single digit number (often one, generally less than three), we have perhaps developed a habit of contacting the most likely knowledgeable/responsible party directly. Clearly, this is not a scalable pattern. If activity increases, or there is otherwise reason to believe that issues raised can/will be addressed by any/all of those contacted (as happens around release) the mailing list will assuredly see more traffic.Rylis Blueshift wrote: * why are the mailing lists dead (well, if you don't count the spam)? They are a very efficient means of communication in contrast to a BBS, and especially developers are fond of them.
-
- Elite
- Posts: 7243
- Joined: Mon Apr 18, 2005 2:40 pm
- Location: LS87, Buenos Aires, República Argentina
Just in case you're checking out SVN to see where you could start, let me say that the Ogre branch has advanced a bit in an invisible corner of my HD.
The current phase's goal on that port is to tie up all the bits in SVN under a special-purpose scenemanager. The objective of the scenemanager is to provide quick handling of a milliard of objects, in preparation for rich environments, filled with particle clouds and lots of traffic.
I have a structure sitting on my HD that handles that pretty well, culls by both location and size, allowing a great degree of LODdability - ie, the more performant graphic system you were hoping for.
The structure needs some tweaking, and probably optimizing (it was a proof of concept, it's a mixture of a Kd-tree and a fibonacci heap, sort of speak), but in general it's usable and I was hoping for some time to dedicate to the project of actually building an Ogre SceneManager from it.
If you feel comfortable with Ogre, or wish to dig into that, it would be a nice project. Two interesting things about it: it's a part of the 0.5.x project that lagged terribly behind schedule, so it could use some pushing, but also it's quite disconnected from the rest of the engine, so you wouldn't have to study the entire codebase.
Another interesting project would be the one Hellcat mentioned, the renewed configuration editor. For that, I can't remember whether I commited it or not, but I have a set of classes specifically designed for XML editing. Think TinyXML, but with editing capabilities. That could help quite a lot, you'd only be concerned with user interface.
If you want to pick that task up, I'd have to finish that which I started a while ago and never finished (and now feel ashamed while I remember), which is the documentation of the many options which are not well understood by many, and sometimes not even on the vegastrike.config (they're left to their defaults most of the time).
Whichever you pick, thanks for helping.
And whatever else, thanks for offering.
The current phase's goal on that port is to tie up all the bits in SVN under a special-purpose scenemanager. The objective of the scenemanager is to provide quick handling of a milliard of objects, in preparation for rich environments, filled with particle clouds and lots of traffic.
I have a structure sitting on my HD that handles that pretty well, culls by both location and size, allowing a great degree of LODdability - ie, the more performant graphic system you were hoping for.
The structure needs some tweaking, and probably optimizing (it was a proof of concept, it's a mixture of a Kd-tree and a fibonacci heap, sort of speak), but in general it's usable and I was hoping for some time to dedicate to the project of actually building an Ogre SceneManager from it.
If you feel comfortable with Ogre, or wish to dig into that, it would be a nice project. Two interesting things about it: it's a part of the 0.5.x project that lagged terribly behind schedule, so it could use some pushing, but also it's quite disconnected from the rest of the engine, so you wouldn't have to study the entire codebase.
Another interesting project would be the one Hellcat mentioned, the renewed configuration editor. For that, I can't remember whether I commited it or not, but I have a set of classes specifically designed for XML editing. Think TinyXML, but with editing capabilities. That could help quite a lot, you'd only be concerned with user interface.
If you want to pick that task up, I'd have to finish that which I started a while ago and never finished (and now feel ashamed while I remember), which is the documentation of the many options which are not well understood by many, and sometimes not even on the vegastrike.config (they're left to their defaults most of the time).
Whichever you pick, thanks for helping.
And whatever else, thanks for offering.
-
- Insys Pilot
- Posts: 2
- Joined: Sat Nov 25, 2006 12:52 pm
- Contact:
Thank you all for your prompt and elaborate responses, I highly appreciate it.
In the main engine discussion thread it was said that the new engine would have to be able to cope with the enormous distances needed by VegaStrike while still keeping up with detail. I'm not sure how CS' portals would fit into there, but what I know is that this is only achievable with enough bitspace. Java3D, for example, uses 256 bit fixed-point coordinates, which is enough for all purposes. I believe OGRE uses floats/doubles (?), which are probably both not enough...
There's more stuff I came up with:
k) there's no title screen, only the loading screens. Perhaps a nice title screen for VS should be shown for a few seconds.
l) no main menu -- you're just thrown into things. A main menu would also solve the title screen issue: show title screen a bit -> show main menu which doesn't need a lot of stuff in memory -> start game (if users has decided to do so) -> pick up the stuff from the current state of affairs.
m) related to the previous issue: in-game MOD manager. It should be possible to activate a MOD just by using in-game GUI.
n) The loading screens are sooo 90s :) That's of course not a programming issue, but at least the gals and guys could wear something more futuristic.
I'm not sure what went wrong with your members-only setup, this usually works fine. We could try it again.
A mail<-->BBS gateway would be best, of course... it would be no good if one half of the development would be discussed on the ML and the other on the BBS.
I'm talking mainly about those pesky OGG files.
It's definitely interesting stuff, and I have worked a little bit with OGRE in the past.
Rylis.
I noticed that a lot of these things were proposed earlier, but my main point was uttering what I'm interested in.hellcatv wrote:it sounds like you have some very good ideas here
I'm too, and also busy with a lot of other stuff, so it is well possible that my first results will come up in a few weeks or even months, but this thread should serve as a good starting point for me nevertheless.and we could certainly use another coder, especially at a time like this when folks are busy at school
Ah, that's a good reason of course. I have found CS only to be more mature than OGRE, and that's mainly why I asked.I think ogre keeps taking baby steps towards newer graphical features whereas crystal space has to worry about physics and other things we'd end up ignoring (space physics is a lot different than land physics--esp wrt double precision, etc)
In the main engine discussion thread it was said that the new engine would have to be able to cope with the enormous distances needed by VegaStrike while still keeping up with detail. I'm not sure how CS' portals would fit into there, but what I know is that this is only achievable with enough bitspace. Java3D, for example, uses 256 bit fixed-point coordinates, which is enough for all purposes. I believe OGRE uses floats/doubles (?), which are probably both not enough...
I'm *highly* interested in working on the damage model. A lot of interesting RPG possibilities depend on that.a) damage on the hud: that would be quite useful--we'd probably consider a better internal representation of the damage--though SVN does have a mode where it lists damaged systems---the method for figuring these out is quite poor
The starting system of 0.4.3.least Sol is 1/10 normal size---which means the sun is huge.... which system has the scale problem?
Ah, didn't know that. I wonder who would like hud_lag=1 and why...?d) the visible lag I think is an effect that can be disabled :-) in the config file (hud_lag?)
I'm with you.e) better fonts wouldn't be terribly hard---but given that ogre already has these--may as well just work on ogre?
Yes, and also one that might fit into my airtight schedule.f) better joystick support would be awesome--and probably a reasonable first project to familiarize yourself with things...
It fails locating setup.config here. Well, moot point, I'm going to try SVN and try to fix the problem if I still experience it.g) hmm vssetup worked for me--but ya a few folks had it crash on load---I think it had something to do with the -static flag I passed to gcc
The current approach to things (XML with interleaved GUI instructions) seems very hackish too me. But devising a system that's good isn't easy either. Probably XSLT is the most straightforward way to go?Now what I think a great starter project could be is a new cross platform config file editor that lets folks set the XML to various presets, but also lets folks customize individual values in the XML aside from the various presets (so they can fiddle with things individually after selecting a decent preset)
For me the main incentive lies in carrying out missions with a few human wingmen (or being a wingman to a human leader yourself).h) yes multiplayer development would be awesome--right now it's just deathmatch--but with some cargo and upgrades it could be a lot more like privateer--and with some sort of mutual goals it could become cooperative---course it would only be a few lines of code to make a coop multiplayer gauntlet that spawns enemy bots one by one with increasing difficulty---spiritplumber's WC server has bots flying around that respawn ---so using that code would be easy
Yes, it's hard.i) realistic landing---probably needs a huge graphics overhaul so should wait until ogre work is ready
I'm sure everyone who likes Wing Commander IV would agree that those parts of the game make up a lot of tangible gameplay benefit :)j) definitely worth doing at some point---not sure it's a huge tangible gameplay benefit--but it could make folks more interested in seeing the rest of the game
There's more stuff I came up with:
k) there's no title screen, only the loading screens. Perhaps a nice title screen for VS should be shown for a few seconds.
l) no main menu -- you're just thrown into things. A main menu would also solve the title screen issue: show title screen a bit -> show main menu which doesn't need a lot of stuff in memory -> start game (if users has decided to do so) -> pick up the stuff from the current state of affairs.
m) related to the previous issue: in-game MOD manager. It should be possible to activate a MOD just by using in-game GUI.
n) The loading screens are sooo 90s :) That's of course not a programming issue, but at least the gals and guys could wear something more futuristic.
I'm on it.ahh yes--- 0.4.3. is sorely out of date--and it's strongly recommended you try out the svn release---or at least some of the mods to the games' releases--because they have newer engines I think even...
I just took a look at the check-out process every now and then and saw OGG files almost every time :)not sure why large binary files are there---- we can look at that...
BBS is a nice way for casual users IMO, but I like the flexibility of mailing lists -- I can use my favorite editor and save my work on large posts regularly, for example.as for mailing lists---it's dead because of the spam mostly :-/ we turned it to member only postings---- but somehow the spam still gets through... we can definitely coordinate there--and it's a good place to post ideas---we can revive it from the dead if there's interest!
I'm not sure what went wrong with your members-only setup, this usually works fine. We could try it again.
A mail<-->BBS gateway would be best, of course... it would be no good if one half of the development would be discussed on the ML and the other on the BBS.
Well, I have to admit I'm not very fond of that admin stuff either, I'd rather do more coding. But if it's for the greater benefit of the project and attracting more contributors, it needs to be done...because when the developers have an hour of time they'd rather spend it on working the engine rather than cleaning up management stuff that hobbles along well enough as is...not a great policy--but we do what we individually like cus it's a volunteer project (for worse and better)....
Yes, it is bigger. But it's definitely not on the star scale, it's probably not even a cap ship :)klauss wrote:You lost me. The sun is really bigger than your ship. Really. Even in 0.4.3.
That's not stuff for Subversion, that's stuff for an FTP server.The binary files are there for two reasons.
A) They're the game's data files, so they're an important part of the project.
I'm talking mainly about those pesky OGG files.
If you say so, I will. Thanks.Nobody has anything to say there. Most of what has to be said is said on the boards... just the way it is. Feel free to use them (the mailing lists), what you say there will be heared just the same.
Is this information on the Wiki?JackS wrote:See also: our model of Sol.
Thank you.it wasn't very hard to clean up some of the more egregiously out-of-place cruft (evilwm,ethereal,...). Should be a little tidier now.
What's worse is that it makes it harder to participate for new people.[Discussion mostly in private] is not a scalable pattern.
Noted. I'm not very positive that I can get that done in the next time, however. We'll see.klauss wrote:If you feel comfortable with Ogre, or wish to dig into that, it would be a nice project. Two interesting things about it: it's a part of the 0.5.x project that lagged terribly behind schedule, so it could use some pushing, but also it's quite disconnected from the rest of the engine, so you wouldn't have to study the entire codebase.
It's definitely interesting stuff, and I have worked a little bit with OGRE in the past.
Yes, I noticed that. My approach would have been to dig into the code to see what they mean and then write it down somewhere, but it's a lot more effective if you'd do it.If you want to pick that task up, I'd have to finish that which I started a while ago and never finished (and now feel ashamed while I remember), which is the documentation of the many options which are not well understood by many, and sometimes not even on the vegastrike.config (they're left to their defaults most of the time).
Thanks for your time. I'm sure we can figure something out.Whichever you pick, thanks for helping.
And whatever else, thanks for offering.
Rylis.
-
- Minister of Information
- Posts: 1895
- Joined: Fri Jan 31, 2003 9:40 pm
- Location: The land of tenure (and diaper changes)
on floats/doubles and precision:Rylis Blueshift wrote: In the main engine discussion thread it was said that the new engine would have to be able to cope with the enormous distances needed by VegaStrike while still keeping up with detail. I'm not sure how CS' portals would fit into there, but what I know is that this is only achievable with enough bitspace. Java3D, for example, uses 256 bit fixed-point coordinates, which is enough for all purposes. I believe OGRE uses floats/doubles (?), which are probably both not enough...
IEEE 754 32 bit float: 1 bit sign, 8 bits biased exponent, 23 bits mantissa
IEEE 754 64 bit double: 1 bit sign, 11 bits biased exponent, 52 bits mantissa
(there's also a hidden 1 before the binary point in non-denorms)
VS uses meters as its basic unit. This gives:
float:
meter precision lost at 2^24 meters = 16,777.216 Kilometers
cm precision lost at 2^17 meters = 131.072 Kliometers
mm precision lost at 2^14 meters = 16.384 Kilometers
double:
meter precision lost at 2^53 meters = 9,007,199,254,740.992 Km = ~347 light-days
cm precision lost at 2^46 meters = ~65.2 light-hours
mm precision lost at 2^43 meters = ~8.1 light hours
for comparison, when furthest from Sol, Pluto is ~7.376 billion km = ~6.8 light hours, so we're in milimeter precision territory out at pluto's dimmest portion of its orbit with doubles. So, for realistic scaling, I think we're fine. For efficiency, objects can be locally defined with floats, their position kept track of with doubles, and any interactions performed in a space where the zero-point has been translated from their current positions, such that float-local arithmetic will be sufficient.
-
- Just a tourist with a frag'd nav console
- Posts: 3
- Joined: Mon Jan 22, 2007 6:14 pm
- Location: Grenoble, France
Hello everyone,
is there anything I can do to help the ogre port ?
I hate it when all I can do is to whine that the project is taking years, while the one who's doing it just doesn't have the time to work on it...
I know how to handle C++ and OpenGL, and I found the Ogre documentation rather complete and readable...
Let me know where to start to reduce the backlog and I'll be started.
GBX
is there anything I can do to help the ogre port ?
I hate it when all I can do is to whine that the project is taking years, while the one who's doing it just doesn't have the time to work on it...
I know how to handle C++ and OpenGL, and I found the Ogre documentation rather complete and readable...
Let me know where to start to reduce the backlog and I'll be started.
GBX
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
I believe that there is a SVN branch called "ogre_branch":
SVN url:
https://vegastrike.svn.sourceforge.net/ ... vegastrike
The data URL is at the SVN url:
https://vegastrike.svn.sourceforge.net/ ... nk/data5.x
Here are the ViewVC links for those urls:
http://vegastrike.svn.sourceforge.net/v ... egastrike/
http://vegastrike.svn.sourceforge.net/v ... k/data5.x/
Yes, the OGRE port is very important, but sadly, none of the current developers seem to have time to work on it.
Anyway, I would talk to Klauss about the status of the OGRE port--he was the person doing that, and I don't know the specifics of what state it's in right now. According to the log, it hasn't been modified for 4 months. Maybe there's some uncommitted stuff sitting around.
Thanks!
SVN url:
https://vegastrike.svn.sourceforge.net/ ... vegastrike
The data URL is at the SVN url:
https://vegastrike.svn.sourceforge.net/ ... nk/data5.x
Here are the ViewVC links for those urls:
http://vegastrike.svn.sourceforge.net/v ... egastrike/
http://vegastrike.svn.sourceforge.net/v ... k/data5.x/
Yes, the OGRE port is very important, but sadly, none of the current developers seem to have time to work on it.
Anyway, I would talk to Klauss about the status of the OGRE port--he was the person doing that, and I don't know the specifics of what state it's in right now. According to the log, it hasn't been modified for 4 months. Maybe there's some uncommitted stuff sitting around.
Thanks!
-
- Just a tourist with a frag'd nav console
- Posts: 3
- Joined: Mon Jan 22, 2007 6:14 pm
- Location: Grenoble, France
I'm glad to hear that (that I can help I mean...)
Ok, that's fine. At least I won't step on his feet
Thanks for your answer (Klauss too), I was frantically searching where to grab the code I need without the whole bunch... (I'm sure I read it somewhere!!! but which topic? dammit!!!! )
It seems there's a concept to understand before starting to code, so I'm on to read some code, docs, etc...
GBX
Thanks for your answer (Klauss too), I was frantically searching where to grab the code I need without the whole bunch... (I'm sure I read it somewhere!!! but which topic? dammit!!!! )
It seems there's a concept to understand before starting to code, so I'm on to read some code, docs, etc...
GBX
I don't like when things don't go as fast as I want so I try to help...
No one should feel useless, there's always a way to help.
No one should feel useless, there's always a way to help.
-
- Elite
- Posts: 7243
- Joined: Mon Apr 18, 2005 2:40 pm
- Location: LS87, Buenos Aires, República Argentina
Yap.ace123 wrote:Maybe there's some uncommitted stuff sitting around.Thanks!
It'll get commited when it's stable.
I have to make some time to work on it... hm...
You only need src/UserInterface, most of it is there (plus some changes to mesher, but core VS in the rest of the source tree is pretty much intact).gbx wrote:Thanks for your answer (Klauss too), I was frantically searching where to grab the code I need without the whole bunch... (I'm sure I read it somewhere!!! but which topic? dammit!!!!
For reading. For building, you may need much more (you know how tricky it is... one mismatching header and pum).
-
- Just a tourist with a frag'd nav console
- Posts: 3
- Joined: Mon Jan 22, 2007 6:14 pm
- Location: Grenoble, France
Ok, the less I grab, the faster I'll figure it out. Unless I miss some parts...klauss wrote: You only need src/UserInterface, most of it is there (plus some changes to mesher, but core VS in the rest of the source tree is pretty much intact).
For reading.
Only too well...klauss wrote: For building, you may need much more (you know how tricky it is... one mismatching header and pum).
I don't like when things don't go as fast as I want so I try to help...
No one should feel useless, there's always a way to help.
No one should feel useless, there's always a way to help.