DirectX Game dev toolkit: feedback much appreciated
Author
Hello all,
I have published on Sourceforge a Game Development Toolkit. It is named "SxDL" and lives here at
http://sourceforge.net/projects/sxdl/
I am now at a quite stable point where I need some feedback before going to version 1.0.1.
The objective of the tool is to provide a very easy to use game development framework in the same spirit and "philosophy" of SDL and CDXLib, but for DX9 and recent TnL hardware :
- Beginner friendly
- Simple
- Fast and Efficient
At this time, SxDL handles only 2D and 2.5D, although everything is 3D inside. SxDL will provide 3D tools early next year.
At the SourceForge download page, you will find seven very simple mini-games and demos, 13 tutorials ( written in french ) and the reference documentation ( 80 pages or so ) written in ( not so good ) english.
Please give it a try, and most importantly post here to let me know what you think.
Thank you in advance.
Very interesting concepts indeed.
I will have more time this weekend to look at it and will post my comments on Monday.
Mike
I will have more time this weekend to look at it and will post my comments on Monday.
Mike
Ok... hopefully constructive criticism follows.
Firstly... I won''t touch it as long as it''s under the GPL. I support Open Source software and all that, but from a personal perspective, as someone hoping to sell shareware from home, I can''t afford to distribute my code.
Secondly, it seems too Windows-centric. Many of the design decisions make it look difficult to port to OpenGL. And many data-types are Windows-specific. Therefore I doubt it would replace SDL without a lot of API changes.
Thirdly, I know it''s a very minor point, but I don''t like being told to allocate entities with ''new'' but being told I''m not allowed to ''delete'' them. Wrap the allocation in a factory method if you don''t want users to follow the standard C++ allocation and deallocation idiom.
Lastly, it looks a bit too heavy-weight in that there are functions for absolutely everything, such as deciding whether we are in game or user interface mode, random number generation, file operations, hardcoded game-specific inputs such as acceleration and braking, viscosity effects, and more. Personally I''d much rather have these things in an add-on library, or roll my own in some cases. You''ve said that you''ve tried to "keep it simple" but it is a massive and complex library in my opinion. It''s certainly an impressive piece of work though!
[ MSVC Fixes | STL Docs | SDL | Game AI | Sockets | C++ Faq Lite | Boost
Asking Questions | Organising code files | My stuff | Tiny XML | STLPort]
Firstly... I won''t touch it as long as it''s under the GPL. I support Open Source software and all that, but from a personal perspective, as someone hoping to sell shareware from home, I can''t afford to distribute my code.
Secondly, it seems too Windows-centric. Many of the design decisions make it look difficult to port to OpenGL. And many data-types are Windows-specific. Therefore I doubt it would replace SDL without a lot of API changes.
Thirdly, I know it''s a very minor point, but I don''t like being told to allocate entities with ''new'' but being told I''m not allowed to ''delete'' them. Wrap the allocation in a factory method if you don''t want users to follow the standard C++ allocation and deallocation idiom.
Lastly, it looks a bit too heavy-weight in that there are functions for absolutely everything, such as deciding whether we are in game or user interface mode, random number generation, file operations, hardcoded game-specific inputs such as acceleration and braking, viscosity effects, and more. Personally I''d much rather have these things in an add-on library, or roll my own in some cases. You''ve said that you''ve tried to "keep it simple" but it is a massive and complex library in my opinion. It''s certainly an impressive piece of work though!
[ MSVC Fixes | STL Docs | SDL | Game AI | Sockets | C++ Faq Lite | Boost
Asking Questions | Organising code files | My stuff | Tiny XML | STLPort]
Secondly, it seems too Windows-centric. Many of the design decisions make it look difficult to port to OpenGL. And many data-types are Windows-specific. Therefore I doubt it would replace SDL without a lot of API changes.
Of course it's not going to replace SDL: it's DirectX! How can it not be Windows-centric? Different ideas, different uses...
[edited by - Raloth on November 16, 2003 9:14:41 PM]
Of course it's not going to replace SDL: it's DirectX! How can it not be Windows-centric? Different ideas, different uses...
[edited by - Raloth on November 16, 2003 9:14:41 PM]
Actually, in the documentation it clearly states that one of the goals is to replace SDL and CDX. I''m just pointing out that it''s failing on this route.
Author
First, thank you to all who took time to review sxdl.
Kylotan :
point 1 : Sxdl is clearly not a tool for professional developpers, that''s one of the reason why it is free and open.
point 2 : sxdl uses D3DX vectors and matrices at this time. Sxdl will soon use its own math code. This has been identified as a major problem for portability. The TCHAR macro will be poundefed as char. Passing the windows app instance at startup wont be portable. All the rest should be fully portable. Have you noted any other portability potential problems ? The Linux / OGL will start soon, mid december. The only thing that wont be present in the tux version is the equivalent of the DirectShow sample grabber. The Linux team seam to be very confident. I trust them, they have quite a lot of experience with Linux/OGL.
point 3 : An SxdL port to C# / MDX is also planned. new without delete is perfectly fine in C#. So that''s why there is no factory in the C++ version.
point 4 : Good point. A lot of stuff has been added to fullfill feature requests, when people were developping mini-games and testing the framework. I guess I need to review the everything for general consistency.
About SDL:
Nothing will replace SDL.
SxDL is doing totally different things.
I mentionned SDL in the docs trying to explain that SxDL has been designed trying to acheive the same objectives :
-simplicity
- etc... ( as explained in my first post )
Apparently, this causes a lot of confusion and could trigger "religious" wars, therefore I am going quicly fix this section in the docs.
Maybe I should also change the name of the tool ?
Kylotan :
point 1 : Sxdl is clearly not a tool for professional developpers, that''s one of the reason why it is free and open.
point 2 : sxdl uses D3DX vectors and matrices at this time. Sxdl will soon use its own math code. This has been identified as a major problem for portability. The TCHAR macro will be poundefed as char. Passing the windows app instance at startup wont be portable. All the rest should be fully portable. Have you noted any other portability potential problems ? The Linux / OGL will start soon, mid december. The only thing that wont be present in the tux version is the equivalent of the DirectShow sample grabber. The Linux team seam to be very confident. I trust them, they have quite a lot of experience with Linux/OGL.
point 3 : An SxdL port to C# / MDX is also planned. new without delete is perfectly fine in C#. So that''s why there is no factory in the C++ version.
point 4 : Good point. A lot of stuff has been added to fullfill feature requests, when people were developping mini-games and testing the framework. I guess I need to review the everything for general consistency.
About SDL:
Nothing will replace SDL.
SxDL is doing totally different things.
I mentionned SDL in the docs trying to explain that SxDL has been designed trying to acheive the same objectives :
-simplicity
- etc... ( as explained in my first post )
Apparently, this causes a lot of confusion and could trigger "religious" wars, therefore I am going quicly fix this section in the docs.
Maybe I should also change the name of the tool ?
I tried to build a little app with SxDL (similar to the Pastels demo) and everything worked in no time.
Strongly agree with Kylotan: very impressive!
Best of luck with this project.
Strongly agree with Kylotan: very impressive!
Best of luck with this project.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
