If the goal (at least one of many) of this book is to prove that Pascal is as capable as C++ in creating games I think it should be about 3D rather than 2D simply because it's considered to be more advanced. With that said I'm personally not sure if I'd like a book on 3D better than one on 2D.

Also focusing on more than one Pascal dialect might be a bit hazardous. The book will be harder to read since you have to check which dialect the example code use (even if we make it clear it's gonna be harder, though maybe not too much :?). Another problem is the differences in the dialects, if we have an example in FreePascal dialect then the reader may not be able to understand how to convert it to the Delphi dialect. Last point is that if this if gonna be a community written book then it might not be a good idea to force everyone that contributes to learn several new dialects...

Another thought. Would this book be about creating the next Quake and using Pascal to show how to do it or show that Pascal is the only way to do it (i.e. will the focus be on the algorithms or the language)? I know this can't be answered right now but it must be know before anyone starts to write.

Oh, and someone seemed to miss:
:arrow: AI for Games using Pascal
:arrow: Create your next Game Scripting Language with Pascal (I'd definitely would want one of these )