OneShot
01-02-2008, 03:36 PM
Started by Dr. Sid
If you are wondering if something is happening .. yes, it is. I'm doing some analysis, mostly for object model and graphic interface. First source code would be laid on water this weekend I guess.
---
Fatty
Sounds great Sid :)
---
Dr.Sid
Ok .. so just you know ..
I have some basic GUI and I have obtained some first models from XabbaRus to test them, so I work on 3D a bit too.
My next micro-goals are to add bitmaps to GUI (it's only frames now), and create 3D viewport, so you can view 3D scene together with the GUI elements. With that I'm trying to load and display Xabba's models with all
details (currently working on textures). I do this in standalone project (for now) and this builds on some of my older testing code.
I guess I'll be able to post some first screens soon showing those models, they look great.
I've just posted 3d viewer with Xabba's Foxtrot into files section. It is actual code of the simulator, showing some basic GUI elements and 3d viewport with the first Xabba's model. Note that all is pretty incomplete. We don't need you to tell us what more we have to do. But we need you to tell us if this does not work at all for you or if you experience some crashes.
Installation is simple .. extract this somewhere, go to bin and run sim.exe.
Just so you know ..
I'm doing some quite minor tweaks in the gui. The problem is: shaders. I have found some texts about shaders in OpenGL and finally I understand why everybody likes them so much. So I too could not help myself and I'm playing with them a little.
Anyway I hope I had enough and that I'll make some more significant stuff this week.
Btw. don't worry, the game will work without shaders and I don't want to put it early versions at all. But it makes many things so easy and good water is practically impossible without them.
---
Zachstar
I just wanted to make a note on this topic.
Shaders are a MUST if you want this game to be used longer than the time it took to develop.
We are just about to get into 2008. Anyone who doesn't have atleast a basic OpenGL shader card has got more problems on his hands than a crappy computer.
Shaders remove so much of the crap involved with coding these games and add so much realism and beauty that it is almost stupid to toss them to the side on even the alpha versions.
Please, Plan to use shaders! It will work fine in the end and make things alot smoother to run.
If you ask me... If you are early enough in development that perhaps your time is better spent basing it off an existing game engine such as OGRE http://www.ogre3d.org/ it is free, Open Source, and can be run on DirectX or OpenGL with shader support, It comes with its own shaders such as HDR Lighting and other effects that are really needed for this subsim.
---
Dr.Sid
Shaders are a MUST if you want this game to be used longer than the time it took to develop. We are just about to get into 2008. Anyone who doesn't have atleast a basic OpenGL shader card has got more problems on his hands than a crappy computer.
Like what ? Shader-less opengl will not be supported anymore or what ?
Shaders remove so much of the crap involved with coding these games and add so much realism and beauty that it is almost stupid to toss them to the side on even the alpha versions.
Flashy graphics simply are not priority of the game. Of ANY game. Game is about gaming. It more important to have multiplayer instead of shaders. That's what I mean by adding shaders later.
If you ask me... If you are early enough in development that perhaps your time is better spent basing it off an existing game engine such as OGRE http://www.ogre3d.org/ it is free, Open Source, and can be run on DirectX or OpenGL with shader support, It comes with its own shaders such as HDR Lighting and other effects that are really needed for this subsim.
Well I'm absolutely not interested in such engines. I'm not doing it so it's done fast or easy or anything, but because it's fun for me to do it.
BTW. HDR needed for subsim ? Where are you from ? nVidia corp ? :-D
---
Zachstar
Like what ? Shader-less opengl will not be supported anymore or what
I do not know. It will be quite bad compared to just about any other 3D game and take longer to develop however.
Flashy graphics simply are not priority of the game. Of ANY game. Game is about gaming. It more important to have multiplayer instead of shaders. That's what I mean by adding shaders later.
Add them in now and save yourself the coding headache. You may be noticing a problem with shadows so far no? Something about them sucking far too much FPS? Shadows work best with a simple shader. Less fuss less work.
Well I'm absolutely not interested in such engines. I'm not doing it so it's done fast or easy or anything, but because it's fun for me to do it.
I understand such. Too bad about the choice tho as it can save you alot of headache. Ill put that aside tho.
While you are making this as a hobby project I do hope you keep in mind how others want to help and not be TOO steadfast about things if that MAY help things along a bit. Keep in mind how many projects fail sadly. Dr. Martin of ORBITER is allowing parts of his hobby project to be worked on by others in the OVP project.
BTW. HDR needed for subsim ? Where are you from ? nVidia corp ? :-D
Not much work (A simple shader really) and it gives a good effect? I don't have to be from a graphics corp to say that it is worth it.
The whole point about these shaders is that they are simple and save alot of headaches when you have to do it manually. I suggest you download OGRE and look at the shader folder to see how simple these shaders are. It is more than worth it to require a card from what? 2004? You can pick up a 2006 card for less than 20 bucks these days.
Oh and no I do not work for a graphics corp.
---
Dr.Sid
Add them in now and save yourself the coding headache. You may be noticing a problem with shadows so far no? Something about them sucking far too much FPS? Shadows work best with a simple shader. Less fuss less work.
I see. Sure .. WHEN I'm doing shadows, I'll seriously consider shaders. Or HDR, or that funky reflective water everybody is so mad about these days. But that is not alpha stage. These are nice things, but less important then most of the other things which means will be implemented later.
The whole point about these shaders is that they are simple and save alot of headaches when you have to do it manually. I suggest you download OGRE and look at the shader folder to see how simple these shaders are. It is more than worth it to require a card from what? 2004? You can pick up a 2006 card for less than 20 bucks these days.
At the stage the graphic engine is now (1 simple texture per object) there is no room for shaders, really. Anyway let's discuss what do think shaders make easier. I'm not aware much about such things. I know shaders allows some things which are not possible otherwise (or would be too slow).
---
Zachstar
They make life easier by not making you write a super optimized shadow and light system as well as others because the basic shadowing on another model (water built for waves) costs you 50-75 percent of your performance.
With those out of the way you spend your time on other things such as more advanced physics and realistic weapon actions. Instead of beating your head on the wall trying to figure out how to gain an extra 2 FPS while keeping the shadows not looking like total crap.
Did I mention light without shaders? Nightmare.
GLSL is not a totally new system. My 2004 video card supports it...
All in All requiring GLSL on the client's system is the best way to improve performance, coding effort, bug potential, etc...
I understand you are at pre-alpha. Just please don't get too far before you consider implementing some very basic shaders. Play around and have some fun with it!
---
MaHuJa
I understand you are at pre-alpha. Just please don't get too far before you consider implementing some very basic shaders. Play around and have some fun with it!
I think I'd say it like this:
Learn what they are and how they work, so that
1 - you can make sure adding them in the future won't be a headache
2 - you don't spend a lot of time doing what shaders could do easily.
Some part of the discussion, at least on one side, was about how requiring shaders won't make the game unusable. I have one thing to add about that: Some of the integrated intel graphics cards, like the one on my laptop, really doesn't like some shaders. In one case it was called the "discoball effect". I don't know what shaders in particular it disliked, or if it's shaders in general. At least it runs, and shaders could be 'turned down' so the shaders causing problems wasn't run. (The game was Sword Of The Stars)
I think what the /fuss/ is all about, is that somebody here is hoping that this project will one day succeed to the point of becoming a rival to DW, the clue being: with the quality of graphics required to appeal to the more casual gamer. Because frankly, DW's graphics are easily a turnoff if you're not already looking for its strong points.
---
Zachstar
GLSL is a standard. If an OpenGL game is overwhelming your card when
other games work then more than likely it is bad drivers or the game
is using more shaders than your card supports. (Cityscapes,
fantasyland, etc...)
A game at sea dosent have to use as many shaders as say. Silent Hunter
4 some basic well written ones will save time on development AND
easily beat out Dangerous Waters in graphical performance.
Therefore if the Intel graphics aren't working right then you ought to
dig deeper and see if there any updates/fixes. Or if it is pre-GLSL
then that is sadly a problem that has to be fixed on the hardware
side. Assuming Dr. Sid goes the path of shaders which I recommend fully.
Also may I ask what the real use of a laptop is in this case? I could
hardly imagine effectively using Dangerous Waters on the road! :)
---
Dr.Sid
> A game at sea dosent have to use as many shaders as say. Silent Hunter
> 4 some basic well written ones will save time on development AND
> easily beat out Dangerous Waters in graphical performance.
Well for me subsim is not about graphics. No shadows, no reflections ?
Fine with me as long as it has good sonar, or as I find good partners
to play with. To beat DW in graphic performance is not a priority. It
is matter of my profesional honor, sure :-) But not priority.
> I think I'd say it like this:
>
> Learn what they are and how they work, so that
> 1 - you can make sure adding them in the future won't be a headache
> 2 - you don't spend a lot of time doing what shaders could do easily.
I hope I'm doing exactly that.
> Some part of the discussion, at least on one side, was about how
requiring shaders won't make the game unusable. I have one thing to
add about that: Some of the integrated intel graphics cards, like the
one on my laptop, really doesn't like some shaders. In one case it was
called the "discoball effect". I don't know what shaders in particular
it disliked, or if it's shaders in general. At least it runs, and
shaders could be 'turned down' so the shaders causing problems wasn't
run. (The game was Sword Of The Stars)
There should be 'low fi' mode sure. But that can be simple. No shades,
no reflections. So no extra coding is needed.
> Because frankly, DW's graphics are easily a turnoff if you're not
already looking for its strong points.
Didn't get the last sentence much .. :-/
---
MaHuJa
>> Because frankly, DW's graphics are easily a turnoff if you're not
>> already looking for its strong points.
>
> Didn't get the last sentence much .. :-/
You said it yourself, pretty much...
> Well for me subsim is not about graphics. No shadows, no reflections ?
> Fine with me as long as it has good sonar, or as I find good partners
> to play with.
The same applies to me. But that's what I mean by "looking for its strong
points". I want a good sonar model, and therefore that matters.
However, for someone new to the genre, the accuracy of the sonar model is
something he can't even begin to recognize. To him, an accurate sonar model
actually just makes things complicated. There, the quality of the graphics start
to matter a lot more.
---
Zachstar
> The same applies to me. But that's what I mean by "looking for its
strong points". I want a good sonar model, and therefore that matters.
>
> However, for someone new to the genre, the accuracy of the sonar
model is something he can't even begin to recognize. To him, an
accurate sonar model actually just makes things complicated. There,
the quality of the graphics start to matter a lot more.
>
> M
>
Agreed. Multiplayer opportunities could be harmed if people arent very
interested to continue learning about the advanced operations under
the waves.
However what about that Client/Server model I mentioned earlier. If
sid were to be able to work on the basics of the code and others work
on the graphics it may give very beneficial aspects to the development
and reduce time before alpha/beta.
Just a thought.
---
Dr.Sid
AS I sad .. at the moment there is no special effects and there is no
need for that for a long time. Shadows is special effects. That can be
(and will be) added later.
I played with the shaders (GLSL) a lot. I'm looking forward the moment
there will be time to use them. But it is simple question of
priorities and special effects thus shaders have very very low priority.
Btw. did you played dangerous waters ? I aim to match it in the first
place, then extend it. For the first phase I don't need the shaders.
Now my engine is models with one texture, 1 light, simple fog. The
most advanced technique I plan to use now is extension for better
specular.
I mean I don't do anything which would be implemented better with
shaders at the moment.
Anyway it would be good to have someone who knows GLSL well right
here. What is your experience ?
---
Zachstar
Sadly not enough to assist you well. Most of my knowledge comes from a
research sprint I did when an OVP project in development was having
performance issues with shadows. I quickly discovered why and dug
deeper. It was to my amazement when someone pointed out how an HDR
shader only took less than 50 lines of code. I think it was more like 20.
Now that I think about it... What you want to do perhaps is get a
"Client/Server model". Make the sim 100 percent command line and
implement graphics as a plugin like Dr. Martin is doing for ORBITER.
This could aid you greatly in saving time by letting others work on
the graphical engine.
---
MaHuJa
> Now that I think about it... What you want to do perhaps is get a
> "Client/Server model". Make the sim 100 percent command line and
> implement graphics as a plugin like Dr. Martin is doing for ORBITER.
> This could aid you greatly in saving time by letting others work on
> the graphical engine.
The interfaces in DW are basically plugins. I suspect they all relied on a 3d
renderer in the 'main game' though.
> However what about that Client/Server model I mentioned earlier. If
> sid were to be able to work on the basics of the code and others work
> on the graphics it may give very beneficial aspects to the development
> and reduce time before alpha/beta.
The biggest issue with plugins, the way I know it, is a matter of interface. How
will the plugins interface with the 'core' engine to get what it needs?
How will the graphics plugin interface with the game engine itself be?
What responsibilities are placed where?
Just thinking out loud here:
Modules
'landscape db'
'platform db' - defines (non-graphics) characteristics of each type of ship etc.
(For the record, mimicing DW here is IMO a mistake.)
'core' - keeps track of all ships/subs/etc, provides all the sonar
*calculations* etc.
'graphics db' - stores models, textures, etc. There may be little or no
executable code here, though I guess there should be some that could abstract
archiving the stuff.
'graphics' - code that renders 3d views
'user interface common stuff' - as in stuff that will be common to several
interfaces - such as the sonar waterfall displays.
'user interface for platform x' - which essentially takes the role that
interfaces had for DW.
And then I haven't even begun thinking about sound.
Creating separate modules is the easy part though - the hard part is making data
flow between them, without making them dependent on each others internals, or
create cyclic dependencies, etc. And graphics being a plugin means it can be
replaced - but each must provide the same interface to the user interface
modules, such that each platform UI doesn't require some specific graphics
module. (Imagine OpenGL required for one platform, DX9 for another, and DX10 for
yet another platform...)
---
Zachstar
You have a multitude of languages (Lua, Python, etc... that easily
provide communication)
Also most modern programming languages are able to communicate with
each other just fine. Orbiter OVP still uses C++ for core and plugins.
Sound is handled through OpenAL that alone is a module.
The goal is allowing the core to function as a full simulation all on
its own. This allows for good multiplayer abilities where the clients
send info to the server and the server decides what to do and sends
the resulting info back to the clients for rendering. Puts a big block
on cheating.
If you are wondering if something is happening .. yes, it is. I'm doing some analysis, mostly for object model and graphic interface. First source code would be laid on water this weekend I guess.
---
Fatty
Sounds great Sid :)
---
Dr.Sid
Ok .. so just you know ..
I have some basic GUI and I have obtained some first models from XabbaRus to test them, so I work on 3D a bit too.
My next micro-goals are to add bitmaps to GUI (it's only frames now), and create 3D viewport, so you can view 3D scene together with the GUI elements. With that I'm trying to load and display Xabba's models with all
details (currently working on textures). I do this in standalone project (for now) and this builds on some of my older testing code.
I guess I'll be able to post some first screens soon showing those models, they look great.
I've just posted 3d viewer with Xabba's Foxtrot into files section. It is actual code of the simulator, showing some basic GUI elements and 3d viewport with the first Xabba's model. Note that all is pretty incomplete. We don't need you to tell us what more we have to do. But we need you to tell us if this does not work at all for you or if you experience some crashes.
Installation is simple .. extract this somewhere, go to bin and run sim.exe.
Just so you know ..
I'm doing some quite minor tweaks in the gui. The problem is: shaders. I have found some texts about shaders in OpenGL and finally I understand why everybody likes them so much. So I too could not help myself and I'm playing with them a little.
Anyway I hope I had enough and that I'll make some more significant stuff this week.
Btw. don't worry, the game will work without shaders and I don't want to put it early versions at all. But it makes many things so easy and good water is practically impossible without them.
---
Zachstar
I just wanted to make a note on this topic.
Shaders are a MUST if you want this game to be used longer than the time it took to develop.
We are just about to get into 2008. Anyone who doesn't have atleast a basic OpenGL shader card has got more problems on his hands than a crappy computer.
Shaders remove so much of the crap involved with coding these games and add so much realism and beauty that it is almost stupid to toss them to the side on even the alpha versions.
Please, Plan to use shaders! It will work fine in the end and make things alot smoother to run.
If you ask me... If you are early enough in development that perhaps your time is better spent basing it off an existing game engine such as OGRE http://www.ogre3d.org/ it is free, Open Source, and can be run on DirectX or OpenGL with shader support, It comes with its own shaders such as HDR Lighting and other effects that are really needed for this subsim.
---
Dr.Sid
Shaders are a MUST if you want this game to be used longer than the time it took to develop. We are just about to get into 2008. Anyone who doesn't have atleast a basic OpenGL shader card has got more problems on his hands than a crappy computer.
Like what ? Shader-less opengl will not be supported anymore or what ?
Shaders remove so much of the crap involved with coding these games and add so much realism and beauty that it is almost stupid to toss them to the side on even the alpha versions.
Flashy graphics simply are not priority of the game. Of ANY game. Game is about gaming. It more important to have multiplayer instead of shaders. That's what I mean by adding shaders later.
If you ask me... If you are early enough in development that perhaps your time is better spent basing it off an existing game engine such as OGRE http://www.ogre3d.org/ it is free, Open Source, and can be run on DirectX or OpenGL with shader support, It comes with its own shaders such as HDR Lighting and other effects that are really needed for this subsim.
Well I'm absolutely not interested in such engines. I'm not doing it so it's done fast or easy or anything, but because it's fun for me to do it.
BTW. HDR needed for subsim ? Where are you from ? nVidia corp ? :-D
---
Zachstar
Like what ? Shader-less opengl will not be supported anymore or what
I do not know. It will be quite bad compared to just about any other 3D game and take longer to develop however.
Flashy graphics simply are not priority of the game. Of ANY game. Game is about gaming. It more important to have multiplayer instead of shaders. That's what I mean by adding shaders later.
Add them in now and save yourself the coding headache. You may be noticing a problem with shadows so far no? Something about them sucking far too much FPS? Shadows work best with a simple shader. Less fuss less work.
Well I'm absolutely not interested in such engines. I'm not doing it so it's done fast or easy or anything, but because it's fun for me to do it.
I understand such. Too bad about the choice tho as it can save you alot of headache. Ill put that aside tho.
While you are making this as a hobby project I do hope you keep in mind how others want to help and not be TOO steadfast about things if that MAY help things along a bit. Keep in mind how many projects fail sadly. Dr. Martin of ORBITER is allowing parts of his hobby project to be worked on by others in the OVP project.
BTW. HDR needed for subsim ? Where are you from ? nVidia corp ? :-D
Not much work (A simple shader really) and it gives a good effect? I don't have to be from a graphics corp to say that it is worth it.
The whole point about these shaders is that they are simple and save alot of headaches when you have to do it manually. I suggest you download OGRE and look at the shader folder to see how simple these shaders are. It is more than worth it to require a card from what? 2004? You can pick up a 2006 card for less than 20 bucks these days.
Oh and no I do not work for a graphics corp.
---
Dr.Sid
Add them in now and save yourself the coding headache. You may be noticing a problem with shadows so far no? Something about them sucking far too much FPS? Shadows work best with a simple shader. Less fuss less work.
I see. Sure .. WHEN I'm doing shadows, I'll seriously consider shaders. Or HDR, or that funky reflective water everybody is so mad about these days. But that is not alpha stage. These are nice things, but less important then most of the other things which means will be implemented later.
The whole point about these shaders is that they are simple and save alot of headaches when you have to do it manually. I suggest you download OGRE and look at the shader folder to see how simple these shaders are. It is more than worth it to require a card from what? 2004? You can pick up a 2006 card for less than 20 bucks these days.
At the stage the graphic engine is now (1 simple texture per object) there is no room for shaders, really. Anyway let's discuss what do think shaders make easier. I'm not aware much about such things. I know shaders allows some things which are not possible otherwise (or would be too slow).
---
Zachstar
They make life easier by not making you write a super optimized shadow and light system as well as others because the basic shadowing on another model (water built for waves) costs you 50-75 percent of your performance.
With those out of the way you spend your time on other things such as more advanced physics and realistic weapon actions. Instead of beating your head on the wall trying to figure out how to gain an extra 2 FPS while keeping the shadows not looking like total crap.
Did I mention light without shaders? Nightmare.
GLSL is not a totally new system. My 2004 video card supports it...
All in All requiring GLSL on the client's system is the best way to improve performance, coding effort, bug potential, etc...
I understand you are at pre-alpha. Just please don't get too far before you consider implementing some very basic shaders. Play around and have some fun with it!
---
MaHuJa
I understand you are at pre-alpha. Just please don't get too far before you consider implementing some very basic shaders. Play around and have some fun with it!
I think I'd say it like this:
Learn what they are and how they work, so that
1 - you can make sure adding them in the future won't be a headache
2 - you don't spend a lot of time doing what shaders could do easily.
Some part of the discussion, at least on one side, was about how requiring shaders won't make the game unusable. I have one thing to add about that: Some of the integrated intel graphics cards, like the one on my laptop, really doesn't like some shaders. In one case it was called the "discoball effect". I don't know what shaders in particular it disliked, or if it's shaders in general. At least it runs, and shaders could be 'turned down' so the shaders causing problems wasn't run. (The game was Sword Of The Stars)
I think what the /fuss/ is all about, is that somebody here is hoping that this project will one day succeed to the point of becoming a rival to DW, the clue being: with the quality of graphics required to appeal to the more casual gamer. Because frankly, DW's graphics are easily a turnoff if you're not already looking for its strong points.
---
Zachstar
GLSL is a standard. If an OpenGL game is overwhelming your card when
other games work then more than likely it is bad drivers or the game
is using more shaders than your card supports. (Cityscapes,
fantasyland, etc...)
A game at sea dosent have to use as many shaders as say. Silent Hunter
4 some basic well written ones will save time on development AND
easily beat out Dangerous Waters in graphical performance.
Therefore if the Intel graphics aren't working right then you ought to
dig deeper and see if there any updates/fixes. Or if it is pre-GLSL
then that is sadly a problem that has to be fixed on the hardware
side. Assuming Dr. Sid goes the path of shaders which I recommend fully.
Also may I ask what the real use of a laptop is in this case? I could
hardly imagine effectively using Dangerous Waters on the road! :)
---
Dr.Sid
> A game at sea dosent have to use as many shaders as say. Silent Hunter
> 4 some basic well written ones will save time on development AND
> easily beat out Dangerous Waters in graphical performance.
Well for me subsim is not about graphics. No shadows, no reflections ?
Fine with me as long as it has good sonar, or as I find good partners
to play with. To beat DW in graphic performance is not a priority. It
is matter of my profesional honor, sure :-) But not priority.
> I think I'd say it like this:
>
> Learn what they are and how they work, so that
> 1 - you can make sure adding them in the future won't be a headache
> 2 - you don't spend a lot of time doing what shaders could do easily.
I hope I'm doing exactly that.
> Some part of the discussion, at least on one side, was about how
requiring shaders won't make the game unusable. I have one thing to
add about that: Some of the integrated intel graphics cards, like the
one on my laptop, really doesn't like some shaders. In one case it was
called the "discoball effect". I don't know what shaders in particular
it disliked, or if it's shaders in general. At least it runs, and
shaders could be 'turned down' so the shaders causing problems wasn't
run. (The game was Sword Of The Stars)
There should be 'low fi' mode sure. But that can be simple. No shades,
no reflections. So no extra coding is needed.
> Because frankly, DW's graphics are easily a turnoff if you're not
already looking for its strong points.
Didn't get the last sentence much .. :-/
---
MaHuJa
>> Because frankly, DW's graphics are easily a turnoff if you're not
>> already looking for its strong points.
>
> Didn't get the last sentence much .. :-/
You said it yourself, pretty much...
> Well for me subsim is not about graphics. No shadows, no reflections ?
> Fine with me as long as it has good sonar, or as I find good partners
> to play with.
The same applies to me. But that's what I mean by "looking for its strong
points". I want a good sonar model, and therefore that matters.
However, for someone new to the genre, the accuracy of the sonar model is
something he can't even begin to recognize. To him, an accurate sonar model
actually just makes things complicated. There, the quality of the graphics start
to matter a lot more.
---
Zachstar
> The same applies to me. But that's what I mean by "looking for its
strong points". I want a good sonar model, and therefore that matters.
>
> However, for someone new to the genre, the accuracy of the sonar
model is something he can't even begin to recognize. To him, an
accurate sonar model actually just makes things complicated. There,
the quality of the graphics start to matter a lot more.
>
> M
>
Agreed. Multiplayer opportunities could be harmed if people arent very
interested to continue learning about the advanced operations under
the waves.
However what about that Client/Server model I mentioned earlier. If
sid were to be able to work on the basics of the code and others work
on the graphics it may give very beneficial aspects to the development
and reduce time before alpha/beta.
Just a thought.
---
Dr.Sid
AS I sad .. at the moment there is no special effects and there is no
need for that for a long time. Shadows is special effects. That can be
(and will be) added later.
I played with the shaders (GLSL) a lot. I'm looking forward the moment
there will be time to use them. But it is simple question of
priorities and special effects thus shaders have very very low priority.
Btw. did you played dangerous waters ? I aim to match it in the first
place, then extend it. For the first phase I don't need the shaders.
Now my engine is models with one texture, 1 light, simple fog. The
most advanced technique I plan to use now is extension for better
specular.
I mean I don't do anything which would be implemented better with
shaders at the moment.
Anyway it would be good to have someone who knows GLSL well right
here. What is your experience ?
---
Zachstar
Sadly not enough to assist you well. Most of my knowledge comes from a
research sprint I did when an OVP project in development was having
performance issues with shadows. I quickly discovered why and dug
deeper. It was to my amazement when someone pointed out how an HDR
shader only took less than 50 lines of code. I think it was more like 20.
Now that I think about it... What you want to do perhaps is get a
"Client/Server model". Make the sim 100 percent command line and
implement graphics as a plugin like Dr. Martin is doing for ORBITER.
This could aid you greatly in saving time by letting others work on
the graphical engine.
---
MaHuJa
> Now that I think about it... What you want to do perhaps is get a
> "Client/Server model". Make the sim 100 percent command line and
> implement graphics as a plugin like Dr. Martin is doing for ORBITER.
> This could aid you greatly in saving time by letting others work on
> the graphical engine.
The interfaces in DW are basically plugins. I suspect they all relied on a 3d
renderer in the 'main game' though.
> However what about that Client/Server model I mentioned earlier. If
> sid were to be able to work on the basics of the code and others work
> on the graphics it may give very beneficial aspects to the development
> and reduce time before alpha/beta.
The biggest issue with plugins, the way I know it, is a matter of interface. How
will the plugins interface with the 'core' engine to get what it needs?
How will the graphics plugin interface with the game engine itself be?
What responsibilities are placed where?
Just thinking out loud here:
Modules
'landscape db'
'platform db' - defines (non-graphics) characteristics of each type of ship etc.
(For the record, mimicing DW here is IMO a mistake.)
'core' - keeps track of all ships/subs/etc, provides all the sonar
*calculations* etc.
'graphics db' - stores models, textures, etc. There may be little or no
executable code here, though I guess there should be some that could abstract
archiving the stuff.
'graphics' - code that renders 3d views
'user interface common stuff' - as in stuff that will be common to several
interfaces - such as the sonar waterfall displays.
'user interface for platform x' - which essentially takes the role that
interfaces had for DW.
And then I haven't even begun thinking about sound.
Creating separate modules is the easy part though - the hard part is making data
flow between them, without making them dependent on each others internals, or
create cyclic dependencies, etc. And graphics being a plugin means it can be
replaced - but each must provide the same interface to the user interface
modules, such that each platform UI doesn't require some specific graphics
module. (Imagine OpenGL required for one platform, DX9 for another, and DX10 for
yet another platform...)
---
Zachstar
You have a multitude of languages (Lua, Python, etc... that easily
provide communication)
Also most modern programming languages are able to communicate with
each other just fine. Orbiter OVP still uses C++ for core and plugins.
Sound is handled through OpenAL that alone is a module.
The goal is allowing the core to function as a full simulation all on
its own. This allows for good multiplayer abilities where the clients
send info to the server and the server decides what to do and sends
the resulting info back to the clients for rendering. Puts a big block
on cheating.