[Ti] like hell freezing over ?

[Yuta Hsu]

On Jun 7, 2005, at 1:20 AM, Chris Olson wrote:

> Especially since you'll only be able to get a real Mac for (maybe)  
> a couple more years.  Apple made a complete ass of themselves  
> today.  How long are they going to keep this web page up?
> http://www.apple.com/powermac/performance/
>
> And how long are they going to keep this one up?
> http://www.apple.com/macosx/features/64bit/

Well, if you want a 'real' Mac, then you need to go back to the 68k.   
These PowerPC chips aren't the 'real' Mac, they're some crappy new  
unknown chip that's different, and because it's different we  
shouldn't accept it.  So what if it's from Motorola and Apple worked  
with them on it.  So what it may actually be better for the future.   
Change is bad... Long live the 'real' Mac!!!!  :)

It's certainly true that those pages you mention above will probably  
have to go (or at least be revised).  But as bad as you may think  
Intel's chips are, they've improved their chips and architecture  
greatly recently (multiple-cores and mobile chips especially).  I,  
for one, am excited about the mobile possibilities.  The G5 may be a  
technically superior architecture, but I still can't get a Powerbook  
with it.  I suspect we'll see better performing and cheaper  
Powerbooks with built-in WiMax... which would not be possible with  
the current PPC roadmap.

> So tell me, how are developers building 64-bit applications and  
> testing them on a 32-bit 3.6 GHz P-IV processor?  They aren't even  
> giving the developers a machine with a Pentium D in it with 64-bit  
> extensions.  As you may well be aware, programming assembly for  
> x86-64 is significantly different than for PowerPC.  And  
> programming assembly for Intel's IA64 is different again.  So how  
> you going to make a "fat binary" for that?  What's more, x86-64  
> isn't even real 64-bit.  It's another extension piled on top of the  
> x86 arch.

Yes, but didn't you say that AMD was the one that came out with  
x86-64?  And aren't you the big AMD fan saying that AMD is more  
innovative and a better partner for Apple?  The Pentium D's just came  
out, so of course Apple can't give the developers devkits with  
those.  Will we see dual-core chips in Macs next year?  Probably not,  
but by 2007 we will.

And yes, assembly is quite different between PPC & x86, but how many  
developers program in assembly for the mass market?  How many Mac  
developers are programming in assembly?  I daresay none.  Apple's  
already done the hard part with xCode, so all we have to do as  
developers is program in the 4GLs we know and love (C, C++, etc) and  
let xCode compile the fat/universal binary.

> And we got a translation layer (Rosetta) that doesn't support apps  
> written with vector (AltiVec) code, nor does it support code that  
> requires a G4 or G5 processor (like Final Cut).  So now what?  The  
> PowerPC version of Final cut doesn't run on the Intel Mac with the  
> translation layer.

Um, I saw Photoshop run on Rosetta, plus its plugins.  I also saw  
iPhoto run on Rosetta.  I guess those 2 apps don't use any Altivec?

> Something's really rotten in AppleVille here.  They better start  
> getting their story straight real quick because smoke and mirrors  
> don't fool me very often.

But then, if you perceive it's all smoke and mirrors, does it matter  
that maybe it isn't?  Perception is reality.  :)
Cheers,
Yuta