Some early opinions.

Well I have been tinkering around with OpenGL as of late. Just basic fixed function pipeline stuff via the 4th edition of the OpenGL SuperBible. I will say OpenGL initially seems as if it is easier to remember functions due to the naming conventions compaired to D3D. But the OpenGL SuperBible is rather lack luster IMHO. The author seems to skim over a lot of things that the Luna Book covers in depth for D3D. Thankfully the Luna book covered the topics so I really don't have any issues. The other issue with the book I find is the code samples. First the book uses FreeGlut. I had to go and compile the library. Would have been better if the author used SDL I think. Next some of the pre-compiled Code samples don't even work. I am running a HD 4850 that supports every feature of OGL up to current standards. There are 2 or 3 FBO code samples that don't work at all but the rest do. I have a feeling it is a error on the authors part and there is no Errata for the book stating so.

So the low down:

OpenGL seems easier to remember but the books teaching the api out there are lack luster. I might be better off with copies of the Red/Orange book. The structure of OpenGL code can be a little harder to follow due to its C Nature. This leads to things being broken up into a lot of small function calls to achieve what you want.

D3D seems like it is put together a little better but harder to remember because of unintuitive naming conventions. Things in the D3D api seem to logically flow together. Mostly because it is more object oriented. Resources for learning it are a lot more structured and detailed.

Each has the pros and each has the cons. Not sure which path I will take yet going to continue to tinker around with this OpenGL book a little longer so I can at least compare code that does things the same way. Hopefully I will get to VBO's soon and away from this glVertex crap. Might just have to skip ahead.