diff --git a/code/ASEParser.cpp b/code/ASEParser.cpp
index 280dfdfc3..647c721f8 100644
--- a/code/ASEParser.cpp
+++ b/code/ASEParser.cpp
@@ -1887,7 +1887,7 @@ void Parser::ParseLV3MeshNormalListBlock(ASE::Mesh& sMesh)
 
 	// Allocate enough storage for the normals
 	sMesh.mNormals.resize(sMesh.mFaces.size()*3,aiVector3D( 0.f, 0.f, 0.f ));
-	unsigned int index, faceIdx = 0xffffffff;
+	unsigned int index, faceIdx = UINT_MAX;
 
 	// FIXME: rewrite this and find out how to interpret the normals
 	// correctly. This is crap.
@@ -1897,7 +1897,7 @@ void Parser::ParseLV3MeshNormalListBlock(ASE::Mesh& sMesh)
 	while (true)	{
 		if ('*' == *filePtr)	{
 			++filePtr;
-			if (faceIdx != 0xffffffff && TokenMatch(filePtr,"MESH_VERTEXNORMAL",17))	{
+			if (faceIdx != UINT_MAX && TokenMatch(filePtr,"MESH_VERTEXNORMAL",17))	{
 				aiVector3D vNormal;
 				ParseLV4MeshFloatTriple(&vNormal.x,index);
 				if (faceIdx >=  sMesh.mFaces.size())
diff --git a/code/COBScene.h b/code/COBScene.h
index 8620f564d..1b7f533bf 100644
--- a/code/COBScene.h
+++ b/code/COBScene.h
@@ -71,7 +71,7 @@ struct Face
 /** COB chunk header information */
 struct ChunkInfo
 {
-	enum {NO_SIZE=0xffffffff};
+	enum {NO_SIZE=UINT_MAX};
 
 	ChunkInfo ()
 		:	id        (0)
@@ -218,7 +218,7 @@ struct Material : ChunkInfo
 	};
 
 	Material() : alpha(),exp(),ior(),ka(),ks(1.f),
-		matnum(0xffffffff),
+		matnum(UINT_MAX),
 		shader(FLAT),autofacet(FACETED),
 		autofacet_angle()
 	{}
diff --git a/code/ColladaHelper.h b/code/ColladaHelper.h
index d45647f9b..eca835598 100644
--- a/code/ColladaHelper.h
+++ b/code/ColladaHelper.h
@@ -429,7 +429,7 @@ struct Sampler
 		,	mMirrorU	()
 		,	mMirrorV	()
 		,	mOp			(aiTextureOp_Multiply)
-		,	mUVId		(0xffffffff)
+		,	mUVId		(UINT_MAX)
 		,	mWeighting  (1.f)
 		,	mMixWithPrevious (1.f)
 	{}
@@ -466,7 +466,7 @@ struct Sampler
 	 */
 	std::string mUVChannel;
 
-	/** Resolved UV channel index or 0xffffffff if not known
+	/** Resolved UV channel index or UINT_MAX if not known
 	 */
 	unsigned int mUVId;
 
diff --git a/code/ColladaLoader.cpp b/code/ColladaLoader.cpp
index 7a1ddc169..5059b54b6 100644
--- a/code/ColladaLoader.cpp
+++ b/code/ColladaLoader.cpp
@@ -948,13 +948,14 @@ void ColladaLoader::CreateAnimation( aiScene* pScene, const ColladaParser& pPars
 			}
 
 			// determine which transform step is affected by this channel
-			entry.mTransformIndex = 0xffffffff;
+			entry.mTransformIndex = SIZE_MAX;
 			for( size_t a = 0; a < srcNode->mTransforms.size(); ++a)
 				if( srcNode->mTransforms[a].mID == entry.mTransformId)
 					entry.mTransformIndex = a;
 
-			if( entry.mTransformIndex == 0xffffffff)
+			if( entry.mTransformIndex == SIZE_MAX) {
 				continue;
+			}
 
 			entry.mChannel = &(*cit);
 			entries.push_back( entry);
@@ -1154,12 +1155,12 @@ void ColladaLoader::AddTexture ( Assimp::MaterialHelper& mat, const ColladaParse
 	mat.AddProperty((float*)&sampler.mWeighting , 1,
 		_AI_MATKEY_TEXBLEND_BASE, type, idx);
 
-	// UV source index ... if we didn't resolve the mapping it is actually just 
+	// UV source index ... if we didn't resolve the mapping, it is actually just 
 	// a guess but it works in most cases. We search for the frst occurence of a
 	// number in the channel name. We assume it is the zero-based index into the
 	// UV channel array of all corresponding meshes. It could also be one-based
 	// for some exporters, but we won't care of it unless someone complains about.
-	if (sampler.mUVId != 0xffffffff)
+	if (sampler.mUVId != UINT_MAX)
 		map = sampler.mUVId;
 	else {
 		map = -1;
@@ -1289,16 +1290,6 @@ void ColladaLoader::BuildMaterials( const ColladaParser& pParser, aiScene* pScen
 		aiString name( matIt->first);
 		mat->AddProperty(&name,AI_MATKEY_NAME);
 
-		// MEGA SUPER MONSTER HACK by Alex ... It's all my fault, yes.
-		// We store the reference to the effect in the material and
-		// return ... we'll add the actual material properties later
-		// after we processed all meshes. During mesh processing,
-		// we evaluate vertex input mappings. Afterwards we should be
-		// able to correctly setup source UV channels for textures.
-
-		// ... moved to ColladaLoader::FillMaterials()
-		// *duck*
-
 		// store the material
 		mMaterialIndexByName[matIt->first] = newMats.size();
 		newMats.push_back( std::pair<Collada::Effect*, aiMaterial*>(const_cast<Collada::Effect*>(&effect),mat) );
diff --git a/code/ColladaParser.cpp b/code/ColladaParser.cpp
index c47df2fc5..9ec05fdd3 100644
--- a/code/ColladaParser.cpp
+++ b/code/ColladaParser.cpp
@@ -1968,7 +1968,7 @@ void ColladaParser::ReadPrimitives( Mesh* pMesh, std::vector<InputChannel>& pPer
 	// determine number of indices coming per vertex 
 	// find the offset index for all per-vertex channels
 	size_t numOffsets = 1;
-	size_t perVertexOffset = 0xffffffff; // invalid value
+	size_t perVertexOffset = SIZE_MAX; // invalid value
 	BOOST_FOREACH( const InputChannel& channel, pPerIndexChannels)
 	{
 		numOffsets = std::max( numOffsets, channel.mOffset+1);
diff --git a/code/ComputeUVMappingProcess.cpp b/code/ComputeUVMappingProcess.cpp
index 04231afc9..7ca37765d 100644
--- a/code/ComputeUVMappingProcess.cpp
+++ b/code/ComputeUVMappingProcess.cpp
@@ -98,7 +98,7 @@ inline unsigned int FindEmptyUVChannel (aiMesh* mesh)
 		if (!mesh->mTextureCoords[m])return m;
 	
 	DefaultLogger::get()->error("Unable to compute UV coordinates, no free UV slot found");
-	return 0xffffffff;
+	return UINT_MAX;
 }
 
 // ------------------------------------------------------------------------------------------------
@@ -455,7 +455,7 @@ void ComputeUVMappingProcess::Execute( aiScene* pScene)
 						{
 							aiMesh* mesh = pScene->mMeshes[m];
 							unsigned int outIdx;
-							if ( mesh->mMaterialIndex != i || ( outIdx = FindEmptyUVChannel(mesh) ) == 0xffffffff ||
+							if ( mesh->mMaterialIndex != i || ( outIdx = FindEmptyUVChannel(mesh) ) == UINT_MAX ||
 								!mesh->mNumVertices)
 							{
 								continue;
diff --git a/code/DefaultIOStream.cpp b/code/DefaultIOStream.cpp
index 2457d0cc8..ba0f17d8c 100644
--- a/code/DefaultIOStream.cpp
+++ b/code/DefaultIOStream.cpp
@@ -108,7 +108,7 @@ size_t DefaultIOStream::FileSize() const
 		return 0;
 	}
 	
-	if (0xffffffff == cachedSize) {
+	if (SIZE_MAX == cachedSize) {
 
 		// TODO: Is that really faster if we're already owning a handle to the file?
 #if defined _WIN32 && !defined __GNUC__
diff --git a/code/DefaultIOStream.h b/code/DefaultIOStream.h
index 0be908931..bcb6244d8 100644
--- a/code/DefaultIOStream.h
+++ b/code/DefaultIOStream.h
@@ -110,7 +110,7 @@ private:
 inline DefaultIOStream::DefaultIOStream () : 
 	mFile		(NULL), 
 	mFilename	(""),
-	cachedSize	(0xffffffff)
+	cachedSize	(SIZE_MAX)
 {
 	// empty
 }
@@ -121,7 +121,7 @@ inline DefaultIOStream::DefaultIOStream (FILE* pFile,
 		const std::string &strFilename) :
 	mFile(pFile), 
 	mFilename(strFilename),
-	cachedSize	(0xffffffff)
+	cachedSize	(SIZE_MAX)
 {
 	// empty
 }
diff --git a/code/FileSystemFilter.h b/code/FileSystemFilter.h
index 4db72147b..dccb82225 100644
--- a/code/FileSystemFilter.h
+++ b/code/FileSystemFilter.h
@@ -50,6 +50,10 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #include "ParsingUtils.h"
 namespace Assimp	{
 
+inline bool IsHex(char s) {
+	return (s>='0' && s<='9') || (s>='a' && s<='f') || (s>='A' && s<='F');	
+}
+
 // ---------------------------------------------------------------------------
 /** File system filter  
  */
@@ -226,9 +230,8 @@ private:
 			else if (*it == '%' && in.end() - it > 2) {
 			
 				// Hex sequence in URIs
-				uint32_t tmp;
-				if( 0xffffffff != (tmp = HexOctetToDecimal(&*it))) {
-					*it = (char)tmp;
+				if( IsHex((&*it)[0]) && IsHex((&*it)[1]) ) {
+					*it = HexOctetToDecimal(&*it);
 					it = in.erase(it+1,it+2);
 					--it;
 				}
diff --git a/code/FindInvalidDataProcess.cpp b/code/FindInvalidDataProcess.cpp
index a90396f18..9a0d42ac5 100644
--- a/code/FindInvalidDataProcess.cpp
+++ b/code/FindInvalidDataProcess.cpp
@@ -90,7 +90,7 @@ void UpdateMeshReferences(aiNode* node, const std::vector<unsigned int>& meshMap
 		for (unsigned int a = 0; a < node->mNumMeshes;++a)	{
 
 			register unsigned int ref = node->mMeshes[a];
-			if (0xffffffff != (ref = meshMapping[ref]))	{
+			if (UINT_MAX != (ref = meshMapping[ref]))	{
 				node->mMeshes[out++] = ref;
 			}
 		}
@@ -130,7 +130,7 @@ void FindInvalidDataProcess::Execute( aiScene* pScene)
 				delete pScene->mMeshes[a];
 				AI_DEBUG_INVALIDATE_PTR(pScene->mMeshes[a]);
 
-				meshMapping[a] = 0xffffffff;
+				meshMapping[a] = UINT_MAX;
 				continue;
 			}
 		}
diff --git a/code/IRRLoader.cpp b/code/IRRLoader.cpp
index 4889a6c34..5decba167 100644
--- a/code/IRRLoader.cpp
+++ b/code/IRRLoader.cpp
@@ -243,7 +243,7 @@ void IRRImporter::CopyMaterial(std::vector<aiMaterial*>& materials,
 {
 	if (inmaterials.empty())	{
 		// Do we have a default material? If not we need to create one
-		if (0xffffffff == defMatIdx)
+		if (UINT_MAX == defMatIdx)
 		{
 			defMatIdx = (unsigned int)materials.size();
 			MaterialHelper* mat = new MaterialHelper();
@@ -1397,7 +1397,7 @@ void IRRImporter::InternReadFile( const std::string& pFile,
 	/* Now process our scenegraph recursively: generate final
 	 * meshes and generate animation channels for all nodes.
 	 */
-	unsigned int defMatIdx = 0xffffffff;
+	unsigned int defMatIdx = UINT_MAX;
 	GenerateGraph(root,tempScene->mRootNode, tempScene,
 		batch, meshes, anims, attach, materials, defMatIdx);
 
diff --git a/code/IRRLoader.h b/code/IRRLoader.h
index 253d3c7fc..abb186957 100644
--- a/code/IRRLoader.h
+++ b/code/IRRLoader.h
@@ -279,7 +279,7 @@ private:
 	 *
 	 *  @param materials Receives an output material
 	 *  @param inmaterials List of input materials
-	 *  @param defMatIdx Default material index - 0xffffffff if not there
+	 *  @param defMatIdx Default material index - UINT_MAX if not present
 	 *  @param mesh Mesh to work on
 	 */
 	void CopyMaterial(std::vector<aiMaterial*>&	 materials,
diff --git a/code/LWOFileData.h b/code/LWOFileData.h
index a66bb7268..682835164 100644
--- a/code/LWOFileData.h
+++ b/code/LWOFileData.h
@@ -431,10 +431,10 @@ struct Texture
 	};
 
 	Texture()
-		: mClipIdx(0xffffffff)
+		: mClipIdx(UINT_MAX)
 		, mStrength			(1.0f)
 		, mUVChannelIndex	("unknown")
-		, mRealUVIndex		(0xffffffff)
+		, mRealUVIndex		(UINT_MAX)
 		, enabled			(true)
 		, blendType			(Additive)
 		, bCanUse			(true)
@@ -645,7 +645,7 @@ struct Layer
 	PointList mTempPoints;
 
 	/** Lists for every point the index of another point
-	    that has been copied from *this* point or 0xffffffff if
+	    that has been copied from *this* point or UINT_MAX if
 		no copy of the point has been made */
 	ReferrerList mPointReferrers;
 
diff --git a/code/LWOLoader.cpp b/code/LWOLoader.cpp
index 53600d6c4..a1f998dc4 100644
--- a/code/LWOLoader.cpp
+++ b/code/LWOLoader.cpp
@@ -91,7 +91,7 @@ bool LWOImporter::CanRead( const std::string& pFile, IOSystem* pIOHandler, bool
 void LWOImporter::SetupProperties(const Importer* pImp)
 {
 	configSpeedFlag  = ( 0 != pImp->GetPropertyInteger(AI_CONFIG_FAVOUR_SPEED,0) ? true : false);
-	configLayerIndex = pImp->GetPropertyInteger (AI_CONFIG_IMPORT_LWO_ONE_LAYER_ONLY,0xffffffff); 
+	configLayerIndex = pImp->GetPropertyInteger (AI_CONFIG_IMPORT_LWO_ONE_LAYER_ONLY,UINT_MAX); 
 	configLayerName  = pImp->GetPropertyString  (AI_CONFIG_IMPORT_LWO_ONE_LAYER_ONLY,"");
 }
 
@@ -181,7 +181,7 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 		// The newer lightwave format allows the user to configure the
 		// loader that just one layer is used. If this is the case
 		// we need to check now whether the requested layer has been found.
-		if (0xffffffff != configLayerIndex && configLayerIndex > mLayers->size())
+		if (UINT_MAX != configLayerIndex && configLayerIndex > mLayers->size())
 			throw DeadlyImportError("LWO2: The requested layer was not found");
 
 		if (configLayerName.length() && !hasNamedLayer)	{
@@ -200,7 +200,7 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 	apcNodes. reserve(mLayers->size());
 	apcMeshes.reserve(mLayers->size()*std::min(((unsigned int)mSurfaces->size()/2u), 1u));
 
-	unsigned int iDefaultSurface = 0xffffffff; // index of the default surface
+	unsigned int iDefaultSurface = UINT_MAX; // index of the default surface
 	for (LayerList::iterator lit = mLayers->begin(), lend = mLayers->end();lit != lend;++lit)	{
 		LWO::Layer& layer = *lit;
 		if (layer.skip)
@@ -225,10 +225,10 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 				if (idx >= mTags->size())
 				{
 					DefaultLogger::get()->warn("LWO: Invalid face surface index");
-					idx = 0xffffffff;
+					idx = UINT_MAX;
 				}
-				if(0xffffffff == idx || 0xffffffff == (idx = _mMapping[idx]))	{
-					if (0xffffffff == iDefaultSurface)	{
+				if(UINT_MAX == idx || UINT_MAX == (idx = _mMapping[idx]))	{
+					if (UINT_MAX == iDefaultSurface)	{
 						iDefaultSurface = (unsigned int)mSurfaces->size();
 						mSurfaces->push_back(LWO::Surface());
 						LWO::Surface& surf = mSurfaces->back();
@@ -239,7 +239,7 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 				}
 				pSorted[idx].push_back(i);
 			}
-			if (0xffffffff == iDefaultSurface) {
+			if (UINT_MAX == iDefaultSurface) {
 				pSorted.erase(pSorted.end()-1);
 			}
 			for (unsigned int p = 0,i = 0;i < mSurfaces->size();++i)	{
@@ -268,10 +268,12 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 				unsigned int vVColorIndices[AI_MAX_NUMBER_OF_COLOR_SETS];
 
 #if _DEBUG
-				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_TEXTURECOORDS;++mui )
-					vUVChannelIndices[mui] = 0xffffffff;
-				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_COLOR_SETS;++mui )
-					vVColorIndices[mui] = 0xffffffff;
+				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_TEXTURECOORDS;++mui ) {
+					vUVChannelIndices[mui] = UINT_MAX;
+				}
+				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_COLOR_SETS;++mui ) {
+					vVColorIndices[mui] = UINT_MAX;
+				}
 #endif
 
 				FindUVChannels(_mSurfaces[i],sorted,layer,vUVChannelIndices);
@@ -280,8 +282,9 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 				// allocate storage for UV and CV channels
 				aiVector3D* pvUV[AI_MAX_NUMBER_OF_TEXTURECOORDS];
 				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_TEXTURECOORDS;++mui )	{
-					if (0xffffffff == vUVChannelIndices[mui])
+					if (UINT_MAX == vUVChannelIndices[mui]) {
 						break;
+					}
 					
 					pvUV[mui] = mesh->mTextureCoords[mui] = new aiVector3D[mesh->mNumVertices];
 
@@ -294,7 +297,7 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 		
 				aiColor4D* pvVC[AI_MAX_NUMBER_OF_COLOR_SETS];
 				for (unsigned int mui = 0; mui < AI_MAX_NUMBER_OF_COLOR_SETS;++mui)	{
-					if (0xffffffff == vVColorIndices[mui]) {
+					if (UINT_MAX == vVColorIndices[mui]) {
 						break;
 					}
 					pvVC[mui] = mesh->mColors[mui] = new aiColor4D[mesh->mNumVertices];
@@ -319,8 +322,9 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 
 						// process UV coordinates
 						for (unsigned int w = 0; w < AI_MAX_NUMBER_OF_TEXTURECOORDS;++w)	{
-							if (0xffffffff == vUVChannelIndices[w])
+							if (UINT_MAX == vUVChannelIndices[w]) {
 								break;
+							}
 							aiVector3D*& pp = pvUV[w];
 							const aiVector2D& src = ((aiVector2D*)&layer.mUVChannels[vUVChannelIndices[w]].rawData[0])[idx];
 							pp->x = src.x;
@@ -337,8 +341,9 @@ void LWOImporter::InternReadFile( const std::string& pFile,
 
 						// process vertex colors
 						for (unsigned int w = 0; w < AI_MAX_NUMBER_OF_COLOR_SETS;++w)	{
-							if (0xffffffff == vVColorIndices[w])
+							if (UINT_MAX == vVColorIndices[w]) {
 								break;
+							}
 							*pvVC[w] = ((aiColor4D*)&layer.mVColorChannels[vVColorIndices[w]].rawData[0])[idx];
 
 							// If a RGB color map is explicitly requested delete the
@@ -614,7 +619,7 @@ void LWOImporter::GenerateNodeGraph(std::vector<aiNode*>& apcNodes)
 void LWOImporter::ResolveTags()
 {
 	// --- this function is used for both LWO2 and LWOB
-	mMapping->resize(mTags->size(),0xffffffff);
+	mMapping->resize(mTags->size(), UINT_MAX);
 	for (unsigned int a = 0; a  < mTags->size();++a)	{
 
 		const std::string& c = (*mTags)[a];
@@ -711,7 +716,7 @@ void LWOImporter::LoadLWOPoints(unsigned int length)
 
 		// initialize all point referrers with the default values
 		mCurLayer->mPointReferrers.reserve	( regularSize + (regularSize>>2u) );
-		mCurLayer->mPointReferrers.resize	( regularSize, 0xffffffff );
+		mCurLayer->mPointReferrers.resize	( regularSize, UINT_MAX );
 	}
 	else mCurLayer->mTempPoints.resize( regularSize );
 
@@ -897,7 +902,7 @@ inline void LWOImporter::DoRecursiveVMAPAssignment(VMapEntry* base, unsigned int
 		base->rawData[idx*base->dims+i]= data[i];
 	}
 
-	if (0xffffffff != (i = refList[idx])) {
+	if (UINT_MAX != (i = refList[idx])) {
 		DoRecursiveVMAPAssignment(base,numRead,i,data);
 	}
 }
@@ -905,7 +910,7 @@ inline void LWOImporter::DoRecursiveVMAPAssignment(VMapEntry* base, unsigned int
 // ------------------------------------------------------------------------------------------------
 inline void AddToSingleLinkedList(ReferrerList& refList, unsigned int srcIdx, unsigned int destIdx)
 {
-	if(0xffffffff == refList[srcIdx])	{
+	if(UINT_MAX == refList[srcIdx])	{
 		refList[srcIdx] = destIdx;
 		return;
 	}
@@ -1030,12 +1035,13 @@ void LWOImporter::LoadLWO2VertexMap(unsigned int length, bool perPoly)
 						if (tmp == srcIdx)
 							break;
 					}
-					while ((tmp = refList[tmp]) != 0xffffffff);
-					if (tmp == 0xffffffff)
+					while ((tmp = refList[tmp]) != UINT_MAX);
+					if (tmp == UINT_MAX) {
 						continue;
+					}
 
 					had = true;
-					refList.resize(refList.size()+1, 0xffffffff);
+					refList.resize(refList.size()+1, UINT_MAX);
 						
 					idx = (unsigned int)pointList.size();
 					src.mIndices[i] = (unsigned int)pointList.size();
@@ -1282,7 +1288,7 @@ void LWOImporter::LoadLWO2File()
 
 				// Continue loading this layer or ignore it? Check the layer index property
 				// NOTE: The first layer is the default layer, so the layer index is one-based now
-				if (0xffffffff != configLayerIndex && configLayerIndex != mLayers->size()-1)	{
+				if (UINT_MAX != configLayerIndex && configLayerIndex != mLayers->size()-1)	{
 					skip = true;
 				}
 				else skip = false;
diff --git a/code/LWOLoader.h b/code/LWOLoader.h
index 4faa631d5..60260e11e 100644
--- a/code/LWOLoader.h
+++ b/code/LWOLoader.h
@@ -227,13 +227,13 @@ private:
 		unsigned int& faces,
 		uint16_t*& cursor, 
 		const uint16_t* const end,
-		unsigned int max = 0xffffffff);
+		unsigned int max = UINT_MAX);
 
 	void CountVertsAndFacesLWOB(unsigned int& verts, 
 		unsigned int& faces,
 		LE_NCONST uint16_t*& cursor, 
 		const uint16_t* const end,
-		unsigned int max = 0xffffffff);
+		unsigned int max = UINT_MAX);
 
 	// -------------------------------------------------------------------
 	/** Read vertices and faces in a LWOB/LWO2 file
@@ -246,7 +246,7 @@ private:
 	void CopyFaceIndicesLWOB(LWO::FaceList::iterator& it,
 		LE_NCONST uint16_t*& cursor, 
 		const uint16_t* const end, 
-		unsigned int max = 0xffffffff);
+		unsigned int max = UINT_MAX);
 
 	// -------------------------------------------------------------------
 	/** Resolve the tag and surface lists that have been loaded.
@@ -376,7 +376,7 @@ protected:
 	TagList* mTags;
 
 	/** Mapping table to convert from tag to surface indices.
-	    0xffffffff indicates that a no corresponding surface is available */
+	    UINT_MAX indicates that a no corresponding surface is available */
 	TagMappingTable* mMapping;
 
 	/** Temporary surface list from the file */
diff --git a/code/LWOMaterial.cpp b/code/LWOMaterial.cpp
index 861524ff4..5fc2bd7e0 100644
--- a/code/LWOMaterial.cpp
+++ b/code/LWOMaterial.cpp
@@ -120,7 +120,7 @@ bool LWOImporter::HandleTextures(MaterialHelper* pcMat, const TextureList& in, a
 				break;
 			case LWO::Texture::UV:
 				{
-					if( 0xffffffff == (*it).mRealUVIndex )	{
+					if( UINT_MAX == (*it).mRealUVIndex )	{
 						// We have no UV index for this texture, so we can't display it
 						continue;
 					}
@@ -390,7 +390,7 @@ char LWOImporter::FindUVChannels(LWO::TextureList& list,
 			ret = 1;
 		
 			// got it.
-			if ((*it).mRealUVIndex == 0xffffffff || (*it).mRealUVIndex == next)
+			if ((*it).mRealUVIndex == UINT_MAX || (*it).mRealUVIndex == next)
 			{
 				(*it).mRealUVIndex = next;
 			}
@@ -467,7 +467,7 @@ void LWOImporter::FindUVChannels(LWO::Surface& surf,
 		}
 	}
 	if (extra < AI_MAX_NUMBER_OF_TEXTURECOORDS) {
-		out[extra] = 0xffffffff;
+		out[extra] = UINT_MAX;
 	}
 }
 
@@ -515,7 +515,7 @@ void LWOImporter::FindVCChannels(const LWO::Surface& surf, LWO::SortedRep& sorte
 		}
 	}
 	if (next != AI_MAX_NUMBER_OF_COLOR_SETS) {
-		out[next] = 0xffffffff;
+		out[next] = UINT_MAX;
 	}
 }
 
diff --git a/code/MD2Loader.cpp b/code/MD2Loader.cpp
index a581cdff8..150d8aaa8 100644
--- a/code/MD2Loader.cpp
+++ b/code/MD2Loader.cpp
@@ -111,8 +111,8 @@ void MD2Importer::SetupProperties(const Importer* pImp)
 	// The 
 	// AI_CONFIG_IMPORT_MD2_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MD2_KEYFRAME,0xffffffff);
-	if(0xffffffff == configFrameID){
+	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MD2_KEYFRAME,-1);
+	if(static_cast<unsigned int>(-1) == configFrameID){
 		configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 }
diff --git a/code/MD3Loader.cpp b/code/MD3Loader.cpp
index 5562c9bdd..ef80ee4f7 100644
--- a/code/MD3Loader.cpp
+++ b/code/MD3Loader.cpp
@@ -431,8 +431,8 @@ void MD3Importer::SetupProperties(const Importer* pImp)
 	// The 
 	// AI_CONFIG_IMPORT_MD3_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MD3_KEYFRAME,0xffffffff);
-	if(0xffffffff == configFrameID) {
+	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MD3_KEYFRAME,-1);
+	if(static_cast<unsigned int>(-1) == configFrameID) {
 		configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 
@@ -469,7 +469,7 @@ void MD3Importer::ReadSkin(Q3Shader::SkinData& fill) const
 void MD3Importer::ReadShader(Q3Shader::ShaderData& fill) const
 {
 	// Determine Q3 model name from given path
-	std::string::size_type s = path.find_last_of("\\/",path.length()-2);
+	const std::string::size_type s = path.find_last_of("\\/",path.length()-2);
 	const std::string model_file = path.substr(s+1,path.length()-(s+2));
 
 	// If no specific dir or file is given, use our default search behaviour
@@ -481,7 +481,7 @@ void MD3Importer::ReadShader(Q3Shader::ShaderData& fill) const
 	else {
 		// If the given string specifies a file, load this file.
 		// Otherwise it's a directory.
-		std::string::size_type st = configShaderFile.find_last_of('.');
+		const std::string::size_type st = configShaderFile.find_last_of('.');
 		if (st == std::string::npos) {
 			
 			if(!Q3Shader::LoadShader(fill,configShaderFile + model_file + ".shader",mIOHandler)) {
diff --git a/code/MD5Parser.cpp b/code/MD5Parser.cpp
index e8961aa20..faae9cbc6 100644
--- a/code/MD5Parser.cpp
+++ b/code/MD5Parser.cpp
@@ -349,7 +349,7 @@ MD5AnimParser::MD5AnimParser(SectionList& mSections)
 	DefaultLogger::get()->debug("MD5AnimParser begin");
 
 	fFrameRate = 24.0f;
-	mNumAnimatedComponents = 0xffffffff;
+	mNumAnimatedComponents = UINT_MAX;
 	for (SectionList::const_iterator iter =  mSections.begin(), iterEnd = mSections.end();iter != iterEnd;++iter) {
 		if ((*iter).mName == "hierarchy")	{
 			// "sheath"	0 63 6 
@@ -398,8 +398,9 @@ MD5AnimParser::MD5AnimParser(SectionList& mSections)
 			desc.iIndex = strtoul10((*iter).mGlobalValue.c_str());
 
 			// we do already know how much storage we will presumably need
-			if (0xffffffff != mNumAnimatedComponents)
+			if (UINT_MAX != mNumAnimatedComponents) {
 				desc.mValues.reserve(mNumAnimatedComponents);
+			}
 
 			// now read all elements (continous list of floats)
 			for (ElementList::const_iterator eit = (*iter).mElements.begin(), eitEnd = (*iter).mElements.end(); eit != eitEnd; ++eit){
@@ -418,8 +419,9 @@ MD5AnimParser::MD5AnimParser(SectionList& mSections)
 			mAnimatedBones.reserve(num);
 
 			// try to guess the number of animated components if that element is not given
-			if (0xffffffff == mNumAnimatedComponents)
+			if (UINT_MAX  == mNumAnimatedComponents) {
 				mNumAnimatedComponents = num * 6;
+			}
 		}
 		else if((*iter).mName == "numAnimatedComponents")	{
 			mAnimatedBones.reserve( strtoul10((*iter).mGlobalValue.c_str()));
diff --git a/code/MDCLoader.cpp b/code/MDCLoader.cpp
index 4b8d3bd66..1b35a8322 100644
--- a/code/MDCLoader.cpp
+++ b/code/MDCLoader.cpp
@@ -187,10 +187,8 @@ void MDCImporter::SetupProperties(const Importer* pImp)
 {
 	// The AI_CONFIG_IMPORT_MDC_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	if(0xffffffff == (this->configFrameID = pImp->GetPropertyInteger(
-		AI_CONFIG_IMPORT_MDC_KEYFRAME,0xffffffff)))
-	{
-		this->configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
+	if(static_cast<unsigned int>(-1) == (configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MDC_KEYFRAME,-1))){
+		configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 }
 
@@ -251,7 +249,7 @@ void MDCImporter::InternReadFile(
 		pScene->mMeshes[i] = NULL;
 
 	// now read all surfaces
-	unsigned int iDefaultMatIndex = 0xffffffff;
+	unsigned int iDefaultMatIndex = UINT_MAX;
 	for (unsigned int i = 0, iNum = 0; i < pcHeader->ulNumSurfaces;++i)
 	{
 		if (!pcSurface->ulNumVertices || !pcSurface->ulNumTriangles)continue;
@@ -276,7 +274,7 @@ void MDCImporter::InternReadFile(
 				::strlen(pcShader->ucName),sizeof(pcShader->ucName)) ));
 		}
 		// need to create a default material
-		else if (0xffffffff == iDefaultMatIndex)
+		else if (UINT_MAX == iDefaultMatIndex)
 		{
 			pcMesh->mMaterialIndex = iDefaultMatIndex = (unsigned int)aszShaders.size();
 			aszShaders.push_back(std::string());
diff --git a/code/MDLLoader.cpp b/code/MDLLoader.cpp
index 85bb59b4b..6f84c8240 100644
--- a/code/MDLLoader.cpp
+++ b/code/MDLLoader.cpp
@@ -101,12 +101,12 @@ bool MDLImporter::CanRead( const std::string& pFile, IOSystem* pIOHandler, bool
 // Setup configuration properties
 void MDLImporter::SetupProperties(const Importer* pImp)
 {
-	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MDL_KEYFRAME,0xffffffff);
+	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_MDL_KEYFRAME,-1);
 
 	// The 
 	// AI_CONFIG_IMPORT_MDL_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	if(0xffffffff == configFrameID)	{
+	if(static_cast<unsigned int>(-1) == configFrameID)	{
 		configFrameID =  pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 
@@ -356,8 +356,8 @@ void MDLImporter::InternReadFile_Quake1( )
 		else
 		{
 			szCurrent += sizeof(uint32_t);
-			unsigned int iSkip = i ? 0xffffffff : 0;
-			this->CreateTexture_3DGS_MDL4(szCurrent,pcSkin->group,&iSkip);
+			unsigned int iSkip = i ? UINT_MAX : 0;
+			CreateTexture_3DGS_MDL4(szCurrent,pcSkin->group,&iSkip);
 			szCurrent += iSkip;
 		}
 	}
@@ -552,7 +552,7 @@ void MDLImporter::InternReadFile_3DGS_MDL345( )
 		pcSkin = (BE_NCONST  MDL::Skin*)szCurrent;
 		AI_SWAP4( pcSkin->group);
 		// create one output image
-		unsigned int iSkip = i ? 0xffffffff : 0;
+		unsigned int iSkip = i ? UINT_MAX : 0;
 		if (5 <= iGSFileVersion)
 		{
 			// MDL5 format could contain MIPmaps
@@ -1240,14 +1240,14 @@ void MDLImporter::SortByMaterials_3DGS_MDL7(
 			if (iMatIndex >= iNumMaterials)	{
 				// sometimes MED writes -1, but normally only if there is only
 				// one skin assigned. No warning in this case
-				if(0xffffffff != iMatIndex)
+				if(UINT_MAX != iMatIndex)
 					DefaultLogger::get()->warn("Index overflow in MDL7 material list [#1]");
 				iMatIndex = iNumMaterials-1;
 			}
 			unsigned int iMatIndex2 = groupData.pcFaces[iFace].iMatIndex[1];
 
 			unsigned int iNum = iMatIndex;
-			if (0xffffffff != iMatIndex2 && iMatIndex != iMatIndex2)	{
+			if (UINT_MAX != iMatIndex2 && iMatIndex != iMatIndex2)	{
 				if (iMatIndex2 >= iNumMaterials)	{
 					// sometimes MED writes -1, but normally only if there is only
 					// one skin assigned. No warning in this case
@@ -1447,7 +1447,7 @@ void MDLImporter::InternReadFile_3DGS_MDL7( )
 			groupData.vPositions.resize(iNumVertices);
 			groupData.vNormals.resize(iNumVertices);
 
-			if (sharedData.apcOutBones)groupData.aiBones.resize(iNumVertices,0xffffffff);
+			if (sharedData.apcOutBones)groupData.aiBones.resize(iNumVertices,UINT_MAX);
 
 			// it is also possible that there are 0 UV coordinate sets
 			if (groupInfo.pcGroup->num_stpts){
@@ -1847,7 +1847,7 @@ void MDLImporter::GenerateOutputMeshes_3DGS_MDL7(
 					// iterate through all face indices
 					for (unsigned int c = 0; c < 3;++c)	{
 						unsigned int iBone = groupData.aiBones[ oldFace.mIndices[c] ];
-						if (0xffffffff != iBone)	{
+						if (UINT_MAX != iBone)	{
 							if (iBone >= iNumOutBones)	{
 								DefaultLogger::get()->error("Bone index overflow. "
 									"The bone index of a vertex exceeds the allowed range. ");
diff --git a/code/MDLLoader.h b/code/MDLLoader.h
index ab4197804..42a8c19a9 100644
--- a/code/MDLLoader.h
+++ b/code/MDLLoader.h
@@ -298,7 +298,7 @@ protected:
 	 * \param iType type of the texture data. No DDS or external
 	 * \param piSkip Receive the number of bytes to skip
 	 * \param pcNew Must point to fully initialized data. Width and 
-	 *        height must be set. If pcNew->pcData is set to 0xffffffff,
+	 *        height must be set. If pcNew->pcData is set to UINT_MAX,
 	 *        piSkip will receive the size of the texture, in bytes, but no
 	 *        color data will be read.
 	 */
diff --git a/code/MDLMaterialLoader.cpp b/code/MDLMaterialLoader.cpp
index 1a0ce542c..716d9afb5 100644
--- a/code/MDLMaterialLoader.cpp
+++ b/code/MDLMaterialLoader.cpp
@@ -49,6 +49,7 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #include "MDLDefaultColorMap.h"
 
 using namespace Assimp;
+static aiTexel* const bad_texel = reinterpret_cast<aiTexel*>(SIZE_MAX);
 
 // ------------------------------------------------------------------------------------------------
 // Find a suitable pallette file or take teh default one
@@ -177,14 +178,14 @@ void MDLImporter::CreateTexture_3DGS_MDL4(const unsigned char* szData,
 		return;
 	}
 
-	bool bNoRead = *piSkip == 0xffffffff;
+	const bool bNoRead = *piSkip == UINT_MAX;
 
 	// allocate a new texture object
 	aiTexture* pcNew = new aiTexture();
 	pcNew->mWidth = pcHeader->skinwidth;
 	pcNew->mHeight = pcHeader->skinheight;
 
-	if (bNoRead)pcNew->pcData = (aiTexel*)0xffffffff;
+	if (bNoRead)pcNew->pcData = bad_texel;
 	ParseTextureColorData(szData,iType,piSkip,pcNew);
 
 	// store the texture
@@ -221,9 +222,12 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 	unsigned int* piSkip,
 	aiTexture* pcNew)
 {
+	const bool do_read = bad_texel != pcNew->pcData;
+
 	// allocate storage for the texture image
-	if ((aiTexel*)0xffffffff != pcNew->pcData)
+	if (do_read) {
 		pcNew->pcData = new aiTexel[pcNew->mWidth * pcNew->mHeight];
+	}
 
 	// R5G6B5 format (with or without MIPs)
 	// ****************************************************************
@@ -233,7 +237,7 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 
 		// copy texture data
 		unsigned int i;
-		if ((aiTexel*)0xffffffff != pcNew->pcData) 
+		if (do_read) 
 		{
 			for (i = 0; i < pcNew->mWidth*pcNew->mHeight;++i)
 			{
@@ -264,7 +268,7 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 
 		// copy texture data
 		unsigned int i;
-		if ((aiTexel*)0xffffffff != pcNew->pcData) 
+		if (do_read) 
 		{
 			for (i = 0; i < pcNew->mWidth*pcNew->mHeight;++i)
 			{
@@ -295,7 +299,7 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 
 		// copy texture data
 		unsigned int i;
-		if ((aiTexel*)0xffffffff != pcNew->pcData)
+		if (do_read)
 		{
 			for (i = 0; i < pcNew->mWidth*pcNew->mHeight;++i)
 			{
@@ -326,7 +330,7 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 
 		// copy texture data
 		unsigned int i;
-		if ((aiTexel*)0xffffffff != pcNew->pcData)
+		if (do_read)
 		{
 			for (i = 0; i < pcNew->mWidth*pcNew->mHeight;++i)
 			{
@@ -355,11 +359,11 @@ void MDLImporter::ParseTextureColorData(const unsigned char* szData,
 
 		// copy texture data
 		unsigned int i;
-		if ((aiTexel*)0xffffffff != pcNew->pcData) 
+		if (do_read) 
 		{
 
 			const unsigned char* szColorMap;
-			this->SearchPalette(&szColorMap);
+			SearchPalette(&szColorMap);
 
 			for (i = 0; i < pcNew->mWidth*pcNew->mHeight;++i)
 			{
@@ -388,7 +392,7 @@ void MDLImporter::CreateTexture_3DGS_MDL5(const unsigned char* szData,
 	unsigned int* piSkip)
 {
 	ai_assert(NULL != piSkip);
-	bool bNoRead = *piSkip == 0xffffffff;
+	bool bNoRead = *piSkip == UINT_MAX;
 
 	// allocate a new texture object
 	aiTexture* pcNew = new aiTexture();
@@ -404,12 +408,14 @@ void MDLImporter::CreateTexture_3DGS_MDL5(const unsigned char* szData,
 	AI_SWAP4(pcNew->mHeight); 
 	szData += sizeof(uint32_t);
 
-	if (bNoRead)pcNew->pcData = (aiTexel*)0xffffffff;
+	if (bNoRead) {
+		pcNew->pcData = bad_texel;
+	}
 
-	// this should not occur - at least the docs say it shouldn't
-	// however, you can easily try out what MED does if you have
+	// this should not occur - at least the docs say it shouldn't.
+	// however, one can easily try out what MED does if you have
 	// a model with a DDS texture and export it to MDL5 ...
-	// yes, you're right. It embedds the DDS texture ... :cry:
+	// yeah, it embedds the DDS file.
 	if (6 == iType)
 	{
 		// this is a compressed texture in DDS format
@@ -418,8 +424,7 @@ void MDLImporter::CreateTexture_3DGS_MDL5(const unsigned char* szData,
 
 		if (!bNoRead)
 		{
-			// place a hint and let the application know that it's
-			// a DDS file
+			// place a hint and let the application know that this is a DDS file
 			pcNew->mHeight = 0;
 			pcNew->achFormatHint[0] = 'd';
 			pcNew->achFormatHint[1] = 'd';
@@ -433,8 +438,7 @@ void MDLImporter::CreateTexture_3DGS_MDL5(const unsigned char* szData,
 	else
 	{
 		// parse the color data of the texture
-		ParseTextureColorData(szData,iType,
-			piSkip,pcNew);
+		ParseTextureColorData(szData,iType,piSkip,pcNew);
 	}
 	*piSkip += sizeof(uint32_t) * 2;
 
@@ -565,7 +569,7 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 			pcNew->mHeight = iHeight;
 
 			unsigned int iSkip = 0;
-			this->ParseTextureColorData(szCurrent,iMasked,&iSkip,pcNew);
+			ParseTextureColorData(szCurrent,iMasked,&iSkip,pcNew);
 
 			// skip length of texture data
 			szCurrent += iSkip;
@@ -576,7 +580,7 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 	// texture instead of material colors ... posssible they have
 	// been converted to MDL7 from other formats, such as MDL5
 	aiColor4D clrTexture;
-	if (pcNew)clrTexture = this->ReplaceTextureWithColor(pcNew);
+	if (pcNew)clrTexture = ReplaceTextureWithColor(pcNew);
 	else clrTexture.r = get_qnan();
 	
 	// check whether a material definition is contained in the skin
@@ -637,9 +641,9 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 
 #undef COLOR_MULITPLY_RGB
 
-		// FIX: Take the opacity from the ambient color
-		// the doc says something else, but it is fact that MED exports the
-		// opacity like this .... ARRRGGHH!
+		// FIX: Take the opacity from the ambient color.
+		// The doc say something else, but it is fact that MED exports the
+		// opacity like this .... oh well.
 		clrTemp.r = pcMatIn->Ambient.a;
 		AI_SWAP4(clrTemp.r);  
 		if (is_not_qnan(clrTexture.r)) {
@@ -670,7 +674,7 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 		pcNew = NULL;
 	}
 
-	// if an ASCII effect description (HLSL?) is contained in the file,
+	// If an ASCII effect description (HLSL?) is contained in the file,
 	// we can simply ignore it ...
 	if (iType & AI_MDL7_SKINTYPE_MATERIAL_ASCDEF)
 	{
@@ -683,7 +687,7 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 
 	// If an embedded texture has been loaded setup the corresponding
 	// data structures in the aiScene instance
-	if (pcNew && this->pScene->mNumTextures <= 999)
+	if (pcNew && pScene->mNumTextures <= 999)
 	{
 
 		// place this as diffuse texture
@@ -698,21 +702,22 @@ void MDLImporter::ParseSkinLump_3DGS_MDL7(
 		pcMatOut->AddProperty(&szFile,AI_MATKEY_TEXTURE_DIFFUSE(0));
 
 		// store the texture
-		if (!this->pScene->mNumTextures)
+		if (!pScene->mNumTextures)
 		{
-			this->pScene->mNumTextures = 1;
-			this->pScene->mTextures = new aiTexture*[1];
-			this->pScene->mTextures[0] = pcNew;
+			pScene->mNumTextures = 1;
+			pScene->mTextures = new aiTexture*[1];
+			pScene->mTextures[0] = pcNew;
 		}
 		else
 		{
-			aiTexture** pc = this->pScene->mTextures;
-			this->pScene->mTextures = new aiTexture*[this->pScene->mNumTextures+1];
-			for (unsigned int i = 0; i < this->pScene->mNumTextures;++i)
-				this->pScene->mTextures[i] = pc[i];
+			aiTexture** pc = pScene->mTextures;
+			pScene->mTextures = new aiTexture*[pScene->mNumTextures+1];
+			for (unsigned int i = 0; i < pScene->mNumTextures;++i) {
+				pScene->mTextures[i] = pc[i];
+			}
 
-			this->pScene->mTextures[this->pScene->mNumTextures] = pcNew;
-			this->pScene->mNumTextures++;
+			pScene->mTextures[pScene->mNumTextures] = pcNew;
+			pScene->mNumTextures++;
 			delete[] pc;
 		}
 	}
@@ -730,7 +735,7 @@ void MDLImporter::SkipSkinLump_3DGS_MDL7(
 	unsigned int iHeight)
 {
 	// get the type of the skin
-	unsigned int iMasked = (unsigned int)(iType & 0xF);
+	const unsigned int iMasked = (unsigned int)(iType & 0xF);
 
 	if (0x6 == iMasked)
 	{
@@ -745,15 +750,15 @@ void MDLImporter::SkipSkinLump_3DGS_MDL7(
 	{
 		if (iMasked || !iType || (iType && iWidth && iHeight))
 		{
-			// ParseTextureColorData(..., aiTexture::pcData == 0xffffffff) will simply
+			// ParseTextureColorData(..., aiTexture::pcData == bad_texel) will simply
 			// return the size of the color data in bytes in iSkip
 			unsigned int iSkip = 0;
 
 			aiTexture tex;
-			tex.pcData = reinterpret_cast<aiTexel*>(0xffffffff);
+			tex.pcData = bad_texel;
 			tex.mHeight = iHeight;
 			tex.mWidth = iWidth;
-			this->ParseTextureColorData(szCurrent,iMasked,&iSkip,&tex);
+			ParseTextureColorData(szCurrent,iMasked,&iSkip,&tex);
 
 			// FIX: Important, otherwise the destructor will crash
 			tex.pcData = NULL;
diff --git a/code/MS3DLoader.cpp b/code/MS3DLoader.cpp
index 0b9c7d417..ca5236204 100644
--- a/code/MS3DLoader.cpp
+++ b/code/MS3DLoader.cpp
@@ -227,7 +227,7 @@ void MS3DImporter::InternReadFile( const std::string& pFile,
 		v.bone_id[0] = stream.GetI1(); 
 		v.ref_cnt = stream.GetI1();
 
-		v.bone_id[1] = v.bone_id[2] = v.bone_id[3] = 0xffffffff;
+		v.bone_id[1] = v.bone_id[2] = v.bone_id[3] = UINT_MAX;
 		v.weights[1] = v.weights[2] = v.weights[3] = 0.f;
 		v.weights[0] = 1.f;
 	}
@@ -279,7 +279,7 @@ void MS3DImporter::InternReadFile( const std::string& pFile,
 			t.triangles[i] = stream.GetI2(); 
 		}
 		t.mat = stream.GetI1(); 
-		if (t.mat == 0xffffffff) {
+		if (t.mat == UINT_MAX) {
 			need_default = true;
 		}
 	}
@@ -402,7 +402,7 @@ void MS3DImporter::InternReadFile( const std::string& pFile,
 
 		for (unsigned int i = 0; i < groups.size(); ++i) {
 			TempGroup& g = groups[i];
-			if (g.mat == 0xffffffff) {
+			if (g.mat == UINT_MAX) {
 				g.mat = materials.size()-1;
 			}
 		}
@@ -491,7 +491,7 @@ void MS3DImporter::InternReadFile( const std::string& pFile,
 
 				const TempVertex& v = vertices[t.indices[i]];
 				for(unsigned int a = 0; a < 4; ++a) {
-					if (v.bone_id[a] != 0xffffffff) {
+					if (v.bone_id[a] != UINT_MAX) {
 						if (v.bone_id[a] >= joints.size()) {
 							throw DeadlyImportError("MS3D: Encountered invalid bone index, file is malformed");
 						}
@@ -533,7 +533,7 @@ void MS3DImporter::InternReadFile( const std::string& pFile,
 					const TempVertex& v = vertices[t.indices[i]];
 					for(unsigned int a = 0; a < 4; ++a) {
 						const unsigned int bone = v.bone_id[a];
-						if(bone==0xffffffff){
+						if(bone==UINT_MAX){
 							continue;
 						}
 
diff --git a/code/MaterialSystem.cpp b/code/MaterialSystem.cpp
index e83a5905c..b928ab1bf 100644
--- a/code/MaterialSystem.cpp
+++ b/code/MaterialSystem.cpp
@@ -71,8 +71,8 @@ aiReturn aiGetMaterialProperty(const aiMaterial* pMat,
 
 		if (prop /* just for safety ... */
 			&& 0 == strcmp( prop->mKey.data, pKey ) 
-			&& (0xffffffff == type  || prop->mSemantic == type) /* 0xffffffff is a wildcard, but this is undocumented :-) */ 
-			&& (0xffffffff == index || prop->mIndex == index))
+			&& (UINT_MAX == type  || prop->mSemantic == type) /* UINT_MAX is a wildcard, but this is undocumented :-) */ 
+			&& (UINT_MAX == index || prop->mIndex == index))
 		{
 			*pPropOut = pMat->mProperties[i];
 			return AI_SUCCESS;
@@ -469,7 +469,7 @@ aiReturn MaterialHelper::AddBinaryProperty (const void* pInput,
 	ai_assert (0 != pSizeInBytes);
 
 	// first search the list whether there is already an entry with this key
-	unsigned int iOutIndex = 0xffffffff;
+	unsigned int iOutIndex = UINT_MAX;
 	for (unsigned int i = 0; i < mNumProperties;++i)	{
 		aiMaterialProperty* prop = mProperties[i];
 
@@ -497,7 +497,7 @@ aiReturn MaterialHelper::AddBinaryProperty (const void* pInput,
 	ai_assert ( MAXLEN > pcNew->mKey.length);
 	strcpy( pcNew->mKey.data, pKey );
 
-	if (0xffffffff != iOutIndex)	{
+	if (UINT_MAX != iOutIndex)	{
 		mProperties[iOutIndex] = pcNew;
 		return AI_SUCCESS;
 	}
diff --git a/code/PretransformVertices.cpp b/code/PretransformVertices.cpp
index 9a6a25e0c..829621e53 100644
--- a/code/PretransformVertices.cpp
+++ b/code/PretransformVertices.cpp
@@ -353,7 +353,7 @@ void PretransformVertices::BuildWCSMeshes(std::vector<aiMesh*>& out, aiMesh** in
 	unsigned int numIn, aiNode* node)
 {
 	// NOTE:
-	//  aiMesh::mNumBones store original source mesh, or 0xffffffff if not a copy
+	//  aiMesh::mNumBones store original source mesh, or UINT_MAX if not a copy
 	//  aiMesh::mBones store reference to abs. transform we multiplied with
 
 	// process meshes
@@ -364,7 +364,7 @@ void PretransformVertices::BuildWCSMeshes(std::vector<aiMesh*>& out, aiMesh** in
 		if (!mesh->mBones || *reinterpret_cast<aiMatrix4x4*>(mesh->mBones) == node->mTransformation) {
 			// yes, we can.
 			mesh->mBones = reinterpret_cast<aiBone**> (&node->mTransformation);
-			mesh->mNumBones = 0xffffffff;
+			mesh->mNumBones = UINT_MAX;
 		}
 		else {
 		
diff --git a/code/Q3DLoader.cpp b/code/Q3DLoader.cpp
index a573396a6..15257c019 100644
--- a/code/Q3DLoader.cpp
+++ b/code/Q3DLoader.cpp
@@ -225,7 +225,7 @@ void Q3DImporter::InternReadFile( const std::string& pFile,
 								if (!i && !a)
 									mesh.prevUVIdx = vec.uvindices[a];
 								else if (vec.uvindices[a] != mesh.prevUVIdx)
-									mesh.prevUVIdx = 0xffffffff;
+									mesh.prevUVIdx = UINT_MAX;
 							}
 						}
 					}
diff --git a/code/Q3DLoader.h b/code/Q3DLoader.h
index 897b2b573..3c2b9ac03 100644
--- a/code/Q3DLoader.h
+++ b/code/Q3DLoader.h
@@ -94,7 +94,7 @@ private:
 		Material()
 			:	diffuse			(0.6f,0.6f,0.6f)
 			,	transparency	(0.f)
-			,	texIdx			(0xffffffff)
+			,	texIdx			(UINT_MAX)
 		{}
 
 		aiString name;
diff --git a/code/SMDLoader.cpp b/code/SMDLoader.cpp
index d2f7c607b..5c6622176 100644
--- a/code/SMDLoader.cpp
+++ b/code/SMDLoader.cpp
@@ -86,16 +86,15 @@ void SMDImporter::SetupProperties(const Importer* pImp)
 	// The 
 	// AI_CONFIG_IMPORT_SMD_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_SMD_KEYFRAME,0xffffffff);
-	if(0xffffffff == configFrameID)	{
+	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_SMD_KEYFRAME,-1);
+	if(static_cast<unsigned int>(-1) == configFrameID)	{
 		configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 }
 
 // ------------------------------------------------------------------------------------------------
 // Imports the given file into the given scene structure. 
-void SMDImporter::InternReadFile( 
-	const std::string& pFile, aiScene* pScene, IOSystem* pIOHandler)
+void SMDImporter::InternReadFile( const std::string& pFile, aiScene* pScene, IOSystem* pIOHandler)
 {
 	boost::scoped_ptr<IOStream> file( pIOHandler->Open( pFile, "rb"));
 
@@ -254,7 +253,7 @@ void SMDImporter::CreateOutputMeshes()
 		iFace =  asTriangles.begin();
 		iFace != asTriangles.end();++iFace,++iNum)
 	{
-		if (0xffffffff == (*iFace).iTexture)aaiFaces[(*iFace).iTexture].push_back( 0 );
+		if (UINT_MAX == (*iFace).iTexture)aaiFaces[(*iFace).iTexture].push_back( 0 );
 		else if ((*iFace).iTexture >= aszTextures.size())
 		{
 			DefaultLogger::get()->error("[SMD/VTA] Material index overflow in face");
@@ -354,7 +353,7 @@ void SMDImporter::CreateOutputMeshes()
 				// that the parent of a vertex is 0xffffffff (if the corresponding
 				// entry in the file was unreadable)
 				// ******************************************************************
-				if (fSum < 0.975f && face.avVertices[iVert].iParentNode != 0xffffffff)
+				if (fSum < 0.975f && face.avVertices[iVert].iParentNode != UINT_MAX)
 				{
 					if (face.avVertices[iVert].iParentNode >= asBones.size())
 					{
diff --git a/code/SMDLoader.h b/code/SMDLoader.h
index 0eaa6deac..fe0e436ca 100644
--- a/code/SMDLoader.h
+++ b/code/SMDLoader.h
@@ -69,7 +69,7 @@ namespace SMD	{
 */
 struct Vertex
 {
-	Vertex() : iParentNode(0xffffffff)
+	Vertex() : iParentNode(UINT_MAX)
 	 {}
 
 	//! Vertex position, normal and texture coordinate
@@ -106,7 +106,7 @@ struct Face
 struct Bone
 {
 	//! Default constructor
-	Bone() : iParent(0xffffffff), bIsUsed(false)
+	Bone() : iParent(UINT_MAX), bIsUsed(false)
 	{
 	}
 
diff --git a/code/SceneCombiner.cpp b/code/SceneCombiner.cpp
index e929b953d..b6baada8e 100644
--- a/code/SceneCombiner.cpp
+++ b/code/SceneCombiner.cpp
@@ -243,17 +243,17 @@ void SceneCombiner::MergeScenes(aiScene** _dest, aiScene* master,
 	ai_assert(NULL != _dest);
 
 	// if _dest points to NULL allocate a new scene. Otherwise clear the old and reuse it
-	if (srcList.empty())
-	{
-		if (*_dest)
-		{
+	if (srcList.empty())	{
+		if (*_dest)	{
 			(*_dest)->~aiScene();
 			SceneCombiner::CopySceneFlat(_dest,master);
 		}
 		else *_dest = master;
 		return;
 	}
-	if (*_dest)(*_dest)->~aiScene();
+	if (*_dest) {
+		(*_dest)->~aiScene();
+	}
 	else *_dest = new aiScene();
 
 	aiScene* dest = *_dest;
@@ -265,20 +265,22 @@ void SceneCombiner::MergeScenes(aiScene** _dest, aiScene* master,
 	}
 
 	// this helper array specifies which scenes are duplicates of others
-	std::vector<unsigned int> duplicates(src.size(),0xffffffff);
+	std::vector<unsigned int> duplicates(src.size(),UINT_MAX);
 
 	// this helper array is used as lookup table several times
 	std::vector<unsigned int> offset(src.size());
 
 	// Find duplicate scenes
-	for (unsigned int i = 0; i < src.size();++i)
-	{
-		if (duplicates[i] != i && duplicates[i] != 0xffffffff)continue;
+	for (unsigned int i = 0; i < src.size();++i) {
+		if (duplicates[i] != i && duplicates[i] != UINT_MAX) {
+			continue;
+		}
+			
 		duplicates[i] = i;
-		for ( unsigned int a = i+1; a < src.size(); ++a)
-		{
-			if (src[i].scene == src[a].scene)
+		for ( unsigned int a = i+1; a < src.size(); ++a)	{
+			if (src[i].scene == src[a].scene) {
 				duplicates[a] = i;
+			}
 		}
 	}
 
diff --git a/code/SceneCombiner.h b/code/SceneCombiner.h
index c5bd7ffc2..9dee65603 100644
--- a/code/SceneCombiner.h
+++ b/code/SceneCombiner.h
@@ -76,7 +76,7 @@ struct NodeAttachmentInfo
 		:	node			(NULL)
 		,	attachToNode	(NULL)
 		,	resolved		(false)
-		,	src_idx			(0xffffffff)
+		,	src_idx			(SIZE_MAX)
 	{}
 
 	NodeAttachmentInfo(aiNode* _scene, aiNode* _attachToNode,size_t idx)
diff --git a/code/ScenePreprocessor.cpp b/code/ScenePreprocessor.cpp
index c73c6de36..cb52abbb3 100644
--- a/code/ScenePreprocessor.cpp
+++ b/code/ScenePreprocessor.cpp
@@ -70,11 +70,11 @@ void ScenePreprocessor::ProcessScene ()
 
 		// Check whether there are meshes with at least one set of uv coordinates ... add a dummy texture for them
 		// meshes without texture coordinates receive a boring gray default material.
-		unsigned int mat0 = 0xffffffff, mat1 = 0xffffffff;
+		unsigned int mat0 = UINT_MAX, mat1 = UINT_MAX;
 		for (unsigned int i = 0; i < scene->mNumMeshes;++i) {
 			if (scene->mMeshes[i]->mTextureCoords[0]) {
 
-				if (mat0 == 0xffffffff) {
+				if (mat0 == UINT_MAX) {
 
 					scene->mMaterials[scene->mNumMaterials] = helper = new MaterialHelper();
 					name.Set("$texture.png");
@@ -88,9 +88,8 @@ void ScenePreprocessor::ProcessScene ()
 				}
 				scene->mMeshes[i]->mMaterialIndex = mat0;
 			}
-			else
-			{
-				if (mat1 == 0xffffffff) {
+			else	{
+				if (mat1 == UINT_MAX) {
 
 					scene->mMaterials[scene->mNumMaterials] = helper = new MaterialHelper();
 					aiColor3D clr(0.6f,0.6f,0.6f);
diff --git a/code/SmoothingGroups.h b/code/SmoothingGroups.h
index f6c4b9cdc..42fe25951 100644
--- a/code/SmoothingGroups.h
+++ b/code/SmoothingGroups.h
@@ -61,8 +61,8 @@ struct FaceWithSmoothingGroup
 
 	
 	//! Indices. .3ds is using uint16. However, after
-	//! an unique vrtex set has been generated it might
-	//! be an index becomes > 2^16
+	//! an unique vrtex set has been generated,
+	//! individual index values might exceed 2^16
 	uint32_t mIndices[3];
 
 	//! specifies to which smoothing group the face belongs to
diff --git a/code/SortByPTypeProcess.cpp b/code/SortByPTypeProcess.cpp
index 66fb1eb86..3777a0734 100644
--- a/code/SortByPTypeProcess.cpp
+++ b/code/SortByPTypeProcess.cpp
@@ -92,7 +92,7 @@ void UpdateNodes(const std::vector<unsigned int>& replaceMeshIndex, aiNode* node
 			unsigned int add = node->mMeshes[m]<<2;
 			for (unsigned int i = 0; i < 4;++i)
 			{
-				if (0xffffffff != replaceMeshIndex[add+i])++newSize;
+				if (UINT_MAX != replaceMeshIndex[add+i])++newSize;
 			}
 		}
 		if (!newSize)
@@ -112,7 +112,7 @@ void UpdateNodes(const std::vector<unsigned int>& replaceMeshIndex, aiNode* node
 				unsigned int add = node->mMeshes[m]<<2;
 				for (unsigned int i = 0; i < 4;++i)
 				{
-					if (0xffffffff != replaceMeshIndex[add+i])
+					if (UINT_MAX != replaceMeshIndex[add+i])
 						*newMeshes++ = replaceMeshIndex[add+i];
 				}
 			}
@@ -147,7 +147,7 @@ void SortByPTypeProcess::Execute( aiScene* pScene)
 
 	bool bAnyChanges = false;
 
-	std::vector<unsigned int> replaceMeshIndex(pScene->mNumMeshes*4,0xffffffff);
+	std::vector<unsigned int> replaceMeshIndex(pScene->mNumMeshes*4,UINT_MAX);
 	std::vector<unsigned int>::iterator meshIdx = replaceMeshIndex.begin();
 	for (unsigned int i = 0; i < pScene->mNumMeshes;++i)
 	{
diff --git a/code/SpatialSort.cpp b/code/SpatialSort.cpp
index e5e149945..b8e6fbb84 100644
--- a/code/SpatialSort.cpp
+++ b/code/SpatialSort.cpp
@@ -310,7 +310,7 @@ void SpatialSort::FindIdenticalPositions( const aiVector3D& pPosition,
 // ------------------------------------------------------------------------------------------------
 unsigned int SpatialSort::GenerateMappingTable(std::vector<unsigned int>& fill,float pRadius) const
 {
-	fill.resize(mPositions.size(),0xffffffff);
+	fill.resize(mPositions.size(),UINT_MAX);
 	float dist, maxDist;
 
 	unsigned int t=0;
diff --git a/code/StandardShapes.cpp b/code/StandardShapes.cpp
index fd2a274d3..858b360b7 100644
--- a/code/StandardShapes.cpp
+++ b/code/StandardShapes.cpp
@@ -401,7 +401,7 @@ void StandardShapes::MakeCone(float height,float radius1,
 		std::swap(radius2,radius1);
 		halfHeight = -halfHeight;
 	}
-	else old = 0xffffffff;
+	else old = SIZE_MAX;
 
 	// Use a large epsilon to check whether the cone is pointy
 	if (radius1 < (radius2-radius1)*10e-3f)radius1 = 0.f;
@@ -460,10 +460,10 @@ void StandardShapes::MakeCone(float height,float radius1,
 	}
 
 	// Need to flip face order?
-	if (0xffffffff != old )
-	{
-		for (size_t s = old; s < positions.size();s += 3)
+	if ( SIZE_MAX != old )	{
+		for (size_t s = old; s < positions.size();s += 3) {
 			std::swap(positions[s],positions[s+1]);
+		}
 	}
 }
 
diff --git a/code/StreamReader.h b/code/StreamReader.h
index ec5b87ffc..76f92bc6d 100644
--- a/code/StreamReader.h
+++ b/code/StreamReader.h
@@ -248,11 +248,11 @@ public:
 	/** Setup a temporary read limit
 	 * 
 	 *  @param limit Maximum number of bytes to be read from
-	 *    the beginning of the file. Passing 0xffffffff
+	 *    the beginning of the file. Specifying UINT_MAX
 	 *    resets the limit to the original end of the stream. */
 	void SetReadLimit(unsigned int _limit)	{
 
-		if (0xffffffff == _limit) {
+		if (UINT_MAX == _limit) {
 			limit = end;
 			return;
 		}
diff --git a/code/UnrealLoader.cpp b/code/UnrealLoader.cpp
index 9ddf32712..b53c6beb5 100644
--- a/code/UnrealLoader.cpp
+++ b/code/UnrealLoader.cpp
@@ -92,8 +92,8 @@ void UnrealImporter::SetupProperties(const Importer* pImp)
 	// The 
 	// AI_CONFIG_IMPORT_UNREAL_KEYFRAME option overrides the
 	// AI_CONFIG_IMPORT_GLOBAL_KEYFRAME option.
-	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_UNREAL_KEYFRAME,0xffffffff);
-	if(0xffffffff == configFrameID)	{
+	configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_UNREAL_KEYFRAME,-1);
+	if(static_cast<unsigned int>(-1) == configFrameID)	{
 		configFrameID = pImp->GetPropertyInteger(AI_CONFIG_IMPORT_GLOBAL_KEYFRAME,0);
 	}
 
diff --git a/code/fast_atof.h b/code/fast_atof.h
index 1b64ec97a..90b67725d 100644
--- a/code/fast_atof.h
+++ b/code/fast_atof.h
@@ -111,11 +111,11 @@ inline unsigned int strtoul16( const char* in, const char** out=0)
 
 // ------------------------------------------------------------------------------------
 // Convert just one hex digit
-// Return value is 0xffffffff if the input is not hex
+// Return value is UINT_MAX if the input character is not a hex digit.
 // ------------------------------------------------------------------------------------
 inline unsigned int HexDigitToDecimal(char in)
 {
-	unsigned int out = 0xffffffff;
+	unsigned int out = UINT_MAX;
 	if (in >= '0' && in <= '9')
 		out = in - '0';
 
@@ -125,13 +125,12 @@ inline unsigned int HexDigitToDecimal(char in)
 	else if (in >= 'A' && in <= 'F')
 		out = 10u + in - 'A';
 
-	// return value is 0xffffffff if the input is not a hex digit
+	// return value is UINT_MAX if the input is not a hex digit
 	return out;
 }
 
 // ------------------------------------------------------------------------------------
-// Convert a hex-encoded octet (2 characters processed)
-// Return value is 0xffffffff if the input is not hex
+// Convert a hex-encoded octet (2 characters, i.e. df or 1a).
 // ------------------------------------------------------------------------------------
 inline uint8_t HexOctetToDecimal(const char* in)
 {