Rosetta -- out with a Lion
Funny how everyone is talking about how little Rosetta has to do, how fast it is, and how portable MacOS X is across processors -- ignoring the fact that Apple has decided (now confirmed) to yank it. What is the justification?
One rumor is that Apple would have to now license Rosetta from IBM, which bought Transitive. We all know that Apple is a very poor company and can't negotiate a good deal with a company like IBM, so no wonder that is out!
Rosetta is conceptually no different than what Apple did for 68K apps under PPC. If Apple rolled their own Rosetta, it may not be fast as what Transitive did, but on today's faster Macs, it might be good enough for those who'd be forced to spend thousands on new software. Good thing we have full employment and lots of extra cash to spend in these boom times! But clearly Apple is incapable of writing something like that or they'd at least make the attempt given they are no doubt facing horrific licensing fees for Rosetta from the Evil Big Blue.
But if you buy all of this "Rosetta has to go to keep progress moving forward" talk, you have to wonder what is so awful in MacOS X that it can't maintain PPC compatibility without some serious side effect we don't know about? I mean, what is it about MacOS Xthat holds back progress? It must be something truly terrible! Otherwise it makes it seem like MacOS X isn't so great with binary translation after all, or supporting universal binaries that all of this chit chat about ARM and Intel binaries make it sound like. Given this horrid flaw, I am not sure why Apple would ever consider changing processors again unless backwards compatibility from day 1 is just out. Well, I am sure the App Store can just sell you your apps back with all of the money you have thanks to our fantastic economy.
Curiously I'll bet it cost Apple more to yank Rosetta out of Lion than the licensing costs for Rosetta, but in typical IT fashion, licensing may be a different bucket than development, so even if Dev costs 10 times more, some exec may still say they are saving money if the bucket for licensing is more precious than the development bucket.
You see this with Help Desk costs at other companies---supporting a Help Desk may cost that company many times more than just fixing a little code that drives calls to the Help Desk, but since Help Desk is a separate bucket from IT, spending even 20x the amount on Help Desk may make sense to an exec, depending on the relative size/constraints of those two budgets. Sure it may cost the company as a whole a lot more, but no one does business that way anymore. If you do a great job with your little bucket, you can get fabulous rewards as an exec even if it hurts the company as a whole (not your problem!) So I can imagine spending more money to yank something that is cheaper to keep makes sense to someone. Must be why I am not a rich exec.
Well thats still a lot more work than what Rosetta does.![]()
Funny how everyone is talking about how little Rosetta has to do, how fast it is, and how portable MacOS X is across processors -- ignoring the fact that Apple has decided (now confirmed) to yank it. What is the justification?
One rumor is that Apple would have to now license Rosetta from IBM, which bought Transitive. We all know that Apple is a very poor company and can't negotiate a good deal with a company like IBM, so no wonder that is out!
Rosetta is conceptually no different than what Apple did for 68K apps under PPC. If Apple rolled their own Rosetta, it may not be fast as what Transitive did, but on today's faster Macs, it might be good enough for those who'd be forced to spend thousands on new software. Good thing we have full employment and lots of extra cash to spend in these boom times! But clearly Apple is incapable of writing something like that or they'd at least make the attempt given they are no doubt facing horrific licensing fees for Rosetta from the Evil Big Blue.
But if you buy all of this "Rosetta has to go to keep progress moving forward" talk, you have to wonder what is so awful in MacOS X that it can't maintain PPC compatibility without some serious side effect we don't know about? I mean, what is it about MacOS Xthat holds back progress? It must be something truly terrible! Otherwise it makes it seem like MacOS X isn't so great with binary translation after all, or supporting universal binaries that all of this chit chat about ARM and Intel binaries make it sound like. Given this horrid flaw, I am not sure why Apple would ever consider changing processors again unless backwards compatibility from day 1 is just out. Well, I am sure the App Store can just sell you your apps back with all of the money you have thanks to our fantastic economy.
Curiously I'll bet it cost Apple more to yank Rosetta out of Lion than the licensing costs for Rosetta, but in typical IT fashion, licensing may be a different bucket than development, so even if Dev costs 10 times more, some exec may still say they are saving money if the bucket for licensing is more precious than the development bucket.
You see this with Help Desk costs at other companies---supporting a Help Desk may cost that company many times more than just fixing a little code that drives calls to the Help Desk, but since Help Desk is a separate bucket from IT, spending even 20x the amount on Help Desk may make sense to an exec, depending on the relative size/constraints of those two budgets. Sure it may cost the company as a whole a lot more, but no one does business that way anymore. If you do a great job with your little bucket, you can get fabulous rewards as an exec even if it hurts the company as a whole (not your problem!) So I can imagine spending more money to yank something that is cheaper to keep makes sense to someone. Must be why I am not a rich exec.