3 ms·

Craig Federighi is the software guy, succeeding Bertrand Serlet and Avie Tevanian as VP in charge of software. He's only been at Apple since 2009, though accord

Craig Federighi is the software guy, succeeding Bertrand Serlet and Avie Tevanian as VP in charge of software. He's only been at Apple since 2009, though according to Wikipedia, he did work at NeXT in the 90s leading WebObjects. He's very charismatic and charming in the keynotes, but maybe he deserves some scrutiny for Apple's recent sfotware problems? Though some of it has to be the yearly cycle of releases and the pressure for glitzy features every Summer and Fall.

jrs95 9y ago
I suppose you could ultimately try and place some blame on him but I feel that it's more likely an issue with middle management. Is that his fault to an extent? Perhaps, but it's hard to tell without knowing a lot more than we ever will without being at Apple.
bsaul 9y ago
It could just be a problem of rank. From what i remember from steve jobs decision, ive is in charge of both hardware and software, but i don't think he would be able to code a javascript hello world to save his life. If Craig doesn't have the authority to prioritize hardening the kernel vs the latest gesture du jour, because Ive wants it badly, then no matter how good and charming you are, you're screwed. I mean, it's not hard to see there's something happening with the priorities set inside the company : iphone X is apparently a great new piece of hardware, but iOS and mac OS are seriously lagging. The last batch of real new features was iOS10 with extensions (which honestly didn't change anything in my personal experience), and before that iOS7 with the new UI look. No serious innovation really happened in Apple software since iOS. I'm talking "Richest company in the world" serious. Swift is the only piece of really innovative software coming out of Apple, and from what i understood, it's a very personal initiative of someone who's not even part of the company anymore. And judging at the pace at which the language is evolving, it seems badly under staffed (considering it's supposed to become the main tool for all new apple software). The way i see it (but frankly i don't know), Scott Forstall was the guy with the vision and the history to lead ambitious new software project, and the other top-level guys dumped him (for relational reasons). And now we start seeing the aftermath of that decision.
- andrewjl 9y ago
  Swift, metal, bitcode / LLVM IR, CloudKit, XPC on iOS, all other new frameworks, machine learning on device, differential privacy. Maybe you meant user-facing features, but all the aforementioned help build better software for users.
  - magnetic 9y ago
    APFS too, with what seemed to be a smooth rollout.
    - ghostcluster 9y ago
      The APFS switchover wrecked a lot of people's drives, and AFPS is slower than the filesystem it replaced. I wouldn't hold it up as an example of Apple software initiatives that are shipped fully-baked.
      - magnetic 9y ago
        Do you have source for that? I can't find any reports of widespread problems. The speed wasn't the part I was impressed with, it was the ability to do in place upgrade of a filesystem on millions of devices without any massive amounts of problems. I find that to be a tour de force. While I'd prefer new filesystems to be faster than older ones, I understand that improvements don't always come in the speed dimension: it can also come in the integrity/data loss one. You can run your filesystem without write ordering, cache enabled, and no-op for SYNC/FLUSH, and it will be very fast, but you will risk data loss. It seems reasonable to me to trade speed for integrity.

jrs95 9y ago

I suppose you could ultimately try and place some blame on him but I feel that it's more likely an issue with middle management. Is that his fault to an extent? Perhaps, but it's hard to tell without knowing a lot more than we ever will without being at Apple.

bsaul 9y ago

It could just be a problem of rank. From what i remember from steve jobs decision, ive is in charge of both hardware and software, but i don't think he would be able to code a javascript hello world to save his life. If Craig doesn't have the authority to prioritize hardening the kernel vs the latest gesture du jour, because Ive wants it badly, then no matter how good and charming you are, you're screwed. I mean, it's not hard to see there's something happening with the priorities set inside the company : iphone X is apparently a great new piece of hardware, but iOS and mac OS are seriously lagging. The last batch of real new features was iOS10 with extensions (which honestly didn't change anything in my personal experience), and before that iOS7 with the new UI look. No serious innovation really happened in Apple software since iOS. I'm talking "Richest company in the world" serious. Swift is the only piece of really innovative software coming out of Apple, and from what i understood, it's a very personal initiative of someone who's not even part of the company anymore. And judging at the pace at which the language is evolving, it seems badly under staffed (considering it's supposed to become the main tool for all new apple software). The way i see it (but frankly i don't know), Scott Forstall was the guy with the vision and the history to lead ambitious new software project, and the other top-level guys dumped him (for relational reasons). And now we start seeing the aftermath of that decision.

andrewjl 9y ago
Swift, metal, bitcode / LLVM IR, CloudKit, XPC on iOS, all other new frameworks, machine learning on device, differential privacy. Maybe you meant user-facing features, but all the aforementioned help build better software for users.
- magnetic 9y ago
  APFS too, with what seemed to be a smooth rollout.
  - ghostcluster 9y ago
    The APFS switchover wrecked a lot of people's drives, and AFPS is slower than the filesystem it replaced. I wouldn't hold it up as an example of Apple software initiatives that are shipped fully-baked.
    - magnetic 9y ago
      Do you have source for that? I can't find any reports of widespread problems. The speed wasn't the part I was impressed with, it was the ability to do in place upgrade of a filesystem on millions of devices without any massive amounts of problems. I find that to be a tour de force. While I'd prefer new filesystems to be faster than older ones, I understand that improvements don't always come in the speed dimension: it can also come in the integrity/data loss one. You can run your filesystem without write ordering, cache enabled, and no-op for SYNC/FLUSH, and it will be very fast, but you will risk data loss. It seems reasonable to me to trade speed for integrity.