Slashdot Mirror
How Sony, Intel, and Unix Made Apple's Mac a PC Competitor
smaxp writes In 2007, Sony's supply chain lessons, the network effect from the shift to Intel architecture, and a better OS X for developers combined to renew the Mac's growth. The network effects of the Microsoft Wintel ecosystem that Rappaport explained 20 years ago in the Harvard Business Review are no longer a big advantage. By turning itself into a premium PC company with a proprietary OS, Apple has taken the best of PC ecosystem, but avoided taking on the disadvantages.
You don't need 'a great deal of money' to get into OS X either as user or developer
How dare you bring silly things like facts into Apple bashing. Pay no attention to the fact that Apple has sold an entry-level Mac Mini for $499 for the last 9 years. Even Dell's lowest-end desktops only sell for $100 less.
Pay no attention to the fact that Apple has sold an entry-level Mac Mini for $499 for the last 9 years.
They have sold the entry-level Mac Mini for $499 for 1 week. Before that, it was $599.
It used to be $499, then went up to $599 for a few years, now back to $499. Which is all beside the original point: there is not a high barrier to entry for the Mac. And it has a lot of additional value to a lot of people: simple for the beginner, and an entire open-source UNIX for the advanced user, combined with high-quality parts and great service, a big ecosystem of software and services, and almost no viruses or threats to worry about, and a lot of folks (me included) think life is too short to deal with Windows at home.
E pluribus unum
A brief history of the final days of PPC at Apple:
PPC. Always a day late and a dollar too much. Apple wasn't a big enough customer to justify to IBM to spend more on making foundries and there were always supply problems.
IBM was not concerned with power management at the time, and wanted to build bigger and bigger server class hardware. This was before people actually realized one of the huge costs in building an actual large data center was going to be a major cost compared to hardware and flooring (i.e. the 19 inch racks needed to hold the large iron). It wasn't until Enron emerged from bankruptcy in 2004 and started selling off pieces of itself, culminating with the sale of its last real non-debt instrument asset, Prisma, to Ashmore Energy, that the PG&E contract rate handwriting was on the wall, that energy prices were going to be high in California - where most of the data centers live - for the next 10 years to pay for the long term contracts for natural gas, from Texas, for the power generation plants.
By then, it was far too late for IBM to correct its miscalculation and start producing reasonably power efficient chips in time for Apple.
Apple woke up not too coincidentally when PPC had no viable path for mobile and it's probably one of the best moves Jobs ever made, and in hindsight, most common sense. Surprisingly it took him nearly a decade to shed that inherited weight.
I disagree. Apple had already been in talks with P.A. Semi over the PA6T processor to have a G5 class processor without massive liquid cooling requirements for use in mobile.
The G5 processors from IBM were already looking at massive cooling overhead so that they could survive being overclocked to desktop speeds, and P.A. Semi had the answer Apple needed, but the T.I. foundries were unable to accommodate the necessary feature size shrink to get them to where they needed to be in time.
It was either lose a product cycle (or two), while a willing foundry was being searched out and contracted - which likely meant IBM, at a premium cost, or Intel, which does foundries correctly - or jump ship to Intel. This was at a time Steve was in the middle of his Pancreatic cancer, and it looked like he wouldn't be able to push through to a legacy that would survive his death, without a radical change.
It's a testament to the belief of Apple in the P.A. Semi team that they still bought the company, even though the commitment to an Intel switch, meant that the PA6T and the PWRficient were effectively ruled out. At the time, there were massive problems in the memory bandwidth of ARM processors, and the iPhone was being worked on. So they set the P.A. Semi team, as an "acquihire" rather than a "bring the PPC design in house" play to solving that problem. The Apple CPU still beats the Tegra 4, which is the next closest CPU in memory bandwidth, by about a factor of 4 (8, if you count the 64 bit parts).
So it was a chain of events, and Steve's impending mortality, more than anything else, that killed the PPC at Apple, not that there wasn't a path forward into the mobile marketplace (and Apple had in fact built G4-based iPad prototypes, among other things), and not that Intel was a better path forward onto the supply chain. For Intel, it offered a technology demonstrator opportunity that they needed, because no one was pushing their top end tech until one release cycle behind, and for Steve it was a way to ensure his legacy, while getting back at both IBM and Motorola (it's no mistake that the Intel announcement happened so soon after FreeScale divested themselves of the Intel version of their CodeWarrior product), which he took.
Obviously, my view on some of the details is skewed by where I was in the company at the time; I'm certain other people saw other parts of the elephant, so to speak, but that's roughly how I remember the hallway discussion.
One of the great tragedies, I think, is that there was no Official Apple Historian, with Steve's confidence with regard to secrecy of projects, to document the history of Apple so that we could look at it in clear hindsight.