Adobe’s John Nack took some time Saturday to tell us why Soundbooth, Adobe’s new sound editor, is Intel-only. Disregarding the awesome Parallels Desktop for a minute, Soundbooth is the first major Intel-only Mac app. As such, its release has caused something of a tempest in a teapot on the various Mac forums.
There is a simple reason for the uproar: the Mac faithful is used to getting a ridiculously long usable life out of their Macs. I often see people — Mac professionals even — happily running five year old iBook G3’s with the latest OS X version (try that on Windows). But now, an Adobe app refuses to run on a Power Mac G5 less than three months old.
It’s hard to say what Apple’s opinion of Intel-only Mac apps is going to be. On one hand, when the Intel transition was originally announced, they were incredibly careful to insist that the migration would be to Universal, not to Intel. PowerPC Macs would not become second-class citizens. Then again, I can’t imagine Apple feeling all that bad about users being forced (conveniently by third parties) to upgrade to new Intel Macs.
But there is a potential downside to Apple if Intel-only apps become the norm: the loss of flexibility in architecture choices. The transition to Universal has been about two things: moving everyone over to using gcc, and having every app working on both big and little endian processors. That’s where Linux and most of the open source community is, and their stuff runs on everything — from $50 routers, smart phones and set top boxes to supercomputing clusters with thousands of nodes — with a simple recompile. Apple would’ve loved that flexibility, and came very close to achieving it. If Mac developers only had played along…
They might have taken code from their x86 Win32 audio software that includes assembly, and that's difficult to turn into PowerPC code.
Absolutely. And that's not the only advantage to going Intel only. It's always cheaper to target just one platform instead of two. Perhaps Apple hasn't really made a convincing case for third parties to spend the extra effort for PowerPC compatibility (especially when talking about apps that need a high-end machine). I'm not blaming Adobe one bit.
Sorry for a long and a bit off topic post :)
Sure, the flexibility you mention is very good in theory and Apple probably would not have minded to have it. I am not expert in CPUs but at the moment it seems that there really are not that many possibilities to use the flexibility because there really aren't that many processor architectures alive in the space (notebooks and desktops) Apple mainly is in.
MIPS is dead, the future of SPARC does not look too sure either. In the desktop PPC front Freescale seems quite stagnated, IBM concentrates it's efforts to server products and high volume games console parts.
It is true that in embedded space there are more players, and Apple could be interested in that. Also startups like P.A. Semi have a lot of promise, but it is quite improbable that they could produce a chip that is significantly better in a long run than the x86 giants can. The numbers P.A. Semi have showed are impressive but given the rate industry moves they can't afford to have many delays to have the edge and it will be a tough ride for them anyway, just look at Transmeta.
Having being a bit pessimistic about the situation I must say that I would be very delighted if there were more choice, it's mostly a good thing.