Ok, I removed MyImporter as a project reference and instead choose Browse and selected the MyImporter.dll and it worked. What I don't understand, is either way, when I inspect the reference they are the exact same path and copy local is true. Maybe copy local doesn't work when choosing a C++/CLI project as a reference? I guess ...

I changed it to: public   ref class MyTextureImporter : public ContentImporter<TextureContent^> I still get the same error. I checked all the references, all of them are to 3.1. It does however look like I still have stuff from GS 3.0 even though I uninstalled it. I can't rerun the uninstaller because it is no ...

I'm trying to make a Texture importer using the FreeImage library. I'm following the idea in Shawn's blog here: http://blogs.msdn.com/shawnhar/archive/2007/04/02/any-language-you-like-as-long-as-it-s-at-build-time.aspx I can't seem to get it to use my importer. Here is a basic skeleton, hopefully someone can shed light ...

How much space did you partition with Boot Camp for Windows XP? I'm thinking 60 gigs should allow for many SDKs, VS 2005 and XP.

Any thoughts about maybe open sourcing the importer so the community could submit patches to add support for newer features of the Autodesk SDK?I can imagine it is hard to figure out what you should prioritize.  Maybe opening up some parts to the community would help.

The w coordinate is needed in the projection transform. http://en.wikipedia.org/wiki/Transformation_matrix#Perspective_projectionIf you use a vector3 for the position it will be converted to a float4 with 1 as the w coordinate for the perspective transform.

I read on an MSDN Magazine article (Feb 2002) that md arrays aren't as performant and also in an article by Charles Petzold in Beautiful Code which both mentioned the method call vs instruction as being an issue. Granted, the msdn article is old. Range checks are optimized away from the indexer if the JIT compiler can determine you have ...

I am talking about multidimensional arrays not jagged. What the JIT compiler does aside, just look at the IL that is generated: byte[,] construction: newobj instance void uint8[0...,0...]::.ctor(int32, int32) byte[] construction: newarr uint8 byte[,] setting: call instance void uint8[0...,0...]::Set(int32, int32, ...

I'm not talking about jagged arrays (which actually in .Net are faster), I'm talking about multidimensional arrays. Take a look at the IL that is generated. I know, it is surprising but I'm not making it up. This isn't even counting what the JIT will do or how it is handled on the compact framework. new byte[500,500] ldc.i4 0x1f4ldc.i4 ...

Actually you should not use multidimensional arrays because they do not perform well in .Net. Regular arrays have IL instructions (such as newarr, ldelem, ldelema, ldlen, and stelem) whereas multidimensional require method calls. Not only that, I don't think the JIT compiler optimizes the range check for multidimensional arrays. That math ...