Apple and Intel
Well, I have to say that when I woke up and watched the WWDC this morning, I didn’t expect Apple to announce they were going to be switching their core architecture over from the PowerPC to x86.
So, is this good, or bad? Personally, I think it’s a rather good choice for Apple. IBM has blatently not come up with the goods with respect to the 970 (aka G5); there’s been a lot of problems with it, and quite frankly, IBM seems to have its interests vested elsewhere. Apple did the right thing; IBM had become more of a hinderance than a help, so it makes good business sense to move.
On top of this, they have a good framework to build on. OS X is rather a revolutionary operating system. Not only is it built upon a very, very solid base (in the name of a FreeBSD offshoot), but it also enables for a lot of room for expansion. I wrote to someone not too long ago stating that something along the lines of what Apple are doing now was actually totally feasible. Realistically, the kernel that Apple have written needs to be re-modified so that it runs on x86, and the toolchain rebuilt to accommodate the new assembly, both of which they’ve obviously done quite nicely. Yes, there are other issues besides what I’m saying (it’s obviously not that simple), but that’s the general idea.
Also, because we have the Cocoa and Carbon frameworks to build upon, it makes life a heck of a lot simpler. Apps that are natively coded for X can be transported with only a few minor adjustments. I especially like the “universal binary” idea (both instructions for PowerPC and x86 are included in the binary, and the loader simply loads the correct set depending upon what type of system you’re on).
However, the technology that really stood out for me was the ease and simplicity of the Rosetta translator. That really was something. Basically, it converts the PowerPC instructions into x86 instructions, effectively allowing you to run any PowerPC binary on an x86 architecture. That really was something; it might not be fast enough for 3D/games/etc, but it’s certainly fast enough for Office, Photoshop and the like.
So I think the future looks good for Apple. They’re going to do well, methinks.