Sign in to follow this  
  • entries
    455
  • comments
    639
  • views
    422468

GTL & Future Projects

Sign in to follow this  
_the_phantom_

106 views

So, its around 10:30am, I'm sitting in my Object Implementation lesson at college, generally avoiding writing the trival exercises we have todo, based on the prinicple that (a) java is a pile of shite and (b) I know all of this.

As a by product you lucky lucky people get some ramblings from me on stuff I wouldnt mind doing in the near future.

GTL

So, right now my main focus on GTL. For the most part its pretty stable now an I'm happy with the interface. It still needs a DLL version so Rob Loach can get a .Net interface made for it.

It also needs a C and SDL compatible interface bolted on the front, it'll probably just be a matter of making a small adapter lib to go between the calling code and the C++ GTL code.

After a bit of a convastaion with evolutional the other day I came up with the idea of adding a callback system to GTL, which means that once the image is loaded a user function is called which lets you adjust/access the image before it is returned as an Image object. This idea came about as evolutional wanted to produce a collision map from an image, this means the map can be made as part of the loading process and not as a seperate post processing step.

This infact became a pretty powerfull section of the API, as you can do all kinds of things as you have complete access to the image, including the object which holds the pointer to its data, so if you felt the need you could create a new block of memory, apply some processing to the image and move it to this new block of memory and then adjust the orignal object which held the address so that it points to your new image.
Ofcourse, as the Jedi say, with great power comes great responcibilty and in using this section of the API you could completely break the loading process [grin]

The only other thing I want todo with GTL is add some Aux functionality to make loading textures into OGL/D3D/SDL simply. This will probably exist as an add-on library of a sort. The basic idea is to remove the need to write boiler plate code constantly, something we all hate doing I'm sure [grin]

OGLX

One of the major complaints about OpenGL is the lack of a D3DX like library. This is infact a pretty legitimate issue as D3DX is pretty handy for getting things up and running fast (or so D3D programmers keep telling me). So, with that in mind I'm considering implimenting the same idea but for OpenGL.

The main focus will be the maths functions, as producing them shouldnt be a huge problem and I can follow the D3DX way of doing things pretty closely. They will most likely use SSE1/2/3 as avaible, however I'm still consdering how I'm going todo that, probably a runtime DLL select system of some sort.

Ofcourse, the only problem with DLL is the overhead involved in calling into them. While this isnt a major issue it certainly needs thinking about to some degree.

The other major part of D3DX is the image loading abilities, personally I feel that GTL covers this, it can handle the file formats D3DX can handle (afaik, I'll double check this) and once the extra aux code is added it should have all the short cut stuff in place as well. However, for calarity of code I'll probably use functions to wrap GTL, so as to keep the interface consistant.

The final major consideration is model loading. Now, iirc, D3DX allows you to quickly load an .X model, so something simular is going to be a good idea. Given their open nature a laoder for iD's formats might be helpfull, combined with something for loading .X files as well I guess.

I dare say I've missed other potentially usefull parts, however as this is only my first pondering on it I'm not too worried about that and I'll revisit the topic later I'm sure.

oglFX

Last night on IRC Promit made a comment along the lines of "stop fiddling with texture loaders and write a replacement for cgFX".

At the time I made a rather quick comeback about writing it over Xmas, however while the comment might have been throw away I've considered the idea and the more I think about it the more it seems like a good idea.

Ofcourse, it would primarly be based around GLSL, however the idea its self is a pretty intresting one. I'll need to look into cgFX a bit more to see what kind of undertaking this would turn into, however its currently got my attension.

Who knows, it might even make it into OGLX, a standardised shader framework could certainly be helpfull.

So, those are my immediate plans for the future. As college work isnt really stressing my brain right now I dont see any problem in working the projects in with it, who knows maybe I can develop one into a working 3rd year project for my degree, that would certainly be handy [grin]

As always, comments and feedback are welcome [smile]

Suppose I should pay attension to the class now [grin]
Sign in to follow this  


6 Comments


Recommended Comments

Nice. Since it will be open-source, other developers will be able to work on it and keep improving it's performance and adding new features. I'm totally excited about it!

Share this comment


Link to comment
Dude! Awesome ideas! If I was gonna prioritize, it would be in the same order you wrote about them.

Now some shameless self promotion! You can easily test TGTL visually by writing a simple Thumbnail Extractor for ThumbView! You'll need the normal version, not Lite. If you can't find out how to write your own Thumbnail Extractor by looking around the site; then I've failed and would welcome any input you have. PM me if you have any doubts, or find me on #gamedev @ afternet.org (currently banned for telling a guest that he can change his nick with "/quit yournick", and then he did it eheh).
Heck, if you make such a Extractor I'll put it up on the ThumbView site to give it some free hype ^^

Personally, I'd like to see you work a bit with DigitalDelusion on TGTL. Since he's kinda stopped working on his BeerMap (image lib), but you seem to share some ideas and he has a really nice coding style. Sad to waste good thoughts and planning. Also, he might be able to help you set up a unit test for image testing; as he did one for Beermap.

Share this comment


Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now