Anyone learning OpenGL from code as young as 5 years old (i.e., all fixed-pipeline code) is getting shafted because what they learn is so out of touch with the state of the art in graphics technology. Just like for the Nehe tutorials that whole generations have been brought up on. That's the whole point here, 3d graphics has moved on, and everybody working in it should, too; of course, as with anything, there are always grumpy greybeards who feel that their way of doing things is Good Enough For Them, and therefore should be supported indefinitely.
The arguments on why fixed-pipeline OpenGL should be deprecated have been re-hashed several times convincingly in this thread, I don't have to repeat them here. I don't quite see why the state of the art in graphics should be held back because some people want to see spinning teapots on their iPads without having to learn something new. You seem to be implying that programs written in the past can't be made to work on new platforms; while it's true that it would need changes to the code, please show me the decades-old program in Objective C or Java you mention that you'd like to see working on your iPad without having to change the code.
I hadn't heard of it, but a quick google shows that it's (or rather, was) a closed-source 3d modeler that was discontinued in the late 1990's. That is another situation than the one under discussion; API backwards compatibility doesn't even apply to that product, that product would require a way to run native-code binaries, written for (presumably) Windows (or otherwise, some Unixes), unmodified on the iPad.
The point of my last sentence was that it's impossible, because that long ago Java, iPads, Cocoa etc. didn't even exist. For somebody who wants to port a C++ codebase from those days to a mobile platform, having to re-do the 3d rendering (which is only a very small part of any well-engineered large application anyway) in a programmable pipeline is the least of his worries.
The arguments on why fixed-pipeline OpenGL should be deprecated have been re-hashed several times convincingly in this thread, I don't have to repeat them here. I don't quite see why the state of the art in graphics should be held back because some people want to see spinning teapots on their iPads without having to learn something new. You seem to be implying that programs written in the past can't be made to work on new platforms; while it's true that it would need changes to the code, please show me the decades-old program in Objective C or Java you mention that you'd like to see working on your iPad without having to change the code.