Venturebeat.com: How do you feel about the iPhone platform now?
John Carmack: There are definitely a set of system software things that should improve like multitouch handling. That’s at the top of the list. There are issues with Bluetooth support. One of the spectacular things is that Graeme Divine, who used to work for me at id, is now at Apple in their software division. I have a great man on the inside helping me get questions answered and things accomplished. It will be very good for the evolution of the platform going forward.
From Hell Returns as John Carmack builds Doom Classic for the iPhone
A long time ago, supposedly, Bill Gates realized that Microsoft was making more money from each Mac than from each PC thanks, in large part, to the fact that Mac users used more applications than PC users which, in turn, stemmed from Mac apps being more consistent and easy to use than PC apps. He then — supposedly — wrote a famous letter to John Sculley suggesting that Apple should license its OS to PC manufacturers, and even lined up the meetings.
Sculley and Apple weren’t interested. Microsoft went on to release Windows 3 and Windows 95, eliminating most of the Mac platform’s usability and consistency advantages, and the rest, as they say, is history.
So, the common wisdom is that Apple screwed up a golden opportunity, Microsoft was smart, Apple was stupid, etc.
IBM and Games
But there are two important details missing from this confection: IBM and Gaming.
Microsoft would never have gotten where it is without IBM’s brand name getting it into the enterprise (and then because people used them at work, into the home) — indeed if you read articles from the era Apple’s battle was not against “Microsoft and PC Clones” but IBM — Microsoft and the PC Clones defeated IBM after IBM defeated Apple.
IT has historically hated Apple — at first because the very idea of the personal computer threatened The Priesthood — later because Apple treated users as customers. In 1985 “nobody ever got fired for buying IBM”, and in 1995 “nobody ever got fired for buying Microsoft”.
Back in the 90s, pretty much every large enterprise was not only refusing to buy Apple products, but going to stupid lengths to move Macs out of internal video, marketing, advertising, and desktop publishing sections (where they were both preferred and cost effective). My favorite justification for this was that Apple was a “single vendor” solution (this being from companies that would only buy IBM mainframes, only buy Toshiba laptops, only buy Microsoft operating systems, and only buy Novell servers).
And, of course, Apple was more like IBM than Microsoft – and who benefited from DOS running on non-IBM computers? Certainly not IBM. If IBM had been able to develop its own PC operating system back in 1981, would it have licensed it to Compaq, Dell, and Gateway? If IBM had licensed its OS to clone makers, would IBM still be in the PC business? Would this be considered the genius move that got IBM where it is today (or, say, where it was five years ago when people thought it was in pretty good shape)?
Apple’s other big failure was entirely self-inflicted. At first Apple deliberately spurned games because it wanted the Mac’s graphical interface to be taken seriously as a business tool. But the egregious, repeated, and bewilderingly consistent mishandling of the Mac games market extended far beyond the point of usefulness, if any such point ever existed. In the end, from 1984 to 2006 Apple at best paid lip service to game developers, and generally ignored them or actively made their lives worse.
Apple could easily have ruled the premium home computer market if it had simply paid a little attention to games — or even just refrained from stomping on game developers’ faces. To give some simple examples: Apple steadfastly refused to set joystick standards — let alone sell joysticks — when almost every other computer shipped with joysticks. Apple did its level best to prevent game developers from being able to write full screen programs. When almost every must-have game in the world was using 320×200 256 color graphics, and most Macs were using video hardware that supported this mode, Apple refused to provide system support for this graphics mode — which delayed games like Doom and Quake from appearing on the platform for years, and drove up system requirements to play them at acceptable framerates.
If you cast your mind back to the late 80s and early 90s, when Apple supposedly screwed the pooch by not licensing Mac OS, the home computer market was dominated by Amiga and Atari — both “single vendor” solutions with proprietary operating systems. But both did quite well by actively courting the games market. The Amiga managed to dominate the home computer market for almost five years despite the near criminal incompetence of Commodore’s management. And when Commodore failed, it wasn’t because Commodore didn’t license AmigaDOS to rivals, but because Commodore hadn’t provided a significant upgrade to its most popular model since releasing it. (The Amiga 500 — 1987-1991 — was itself basically just an Amiga 1000 in a cheaper form factor, and its successor, the Amiga 600, was scarcely an upgrade.)
What can we learn from Apple’s failure to become Microsoft in the late 80s?
First, the fact that Microsoft succeeded by licensing MS-DOS and Apple “failed” by not licensing Mac OS doesn’t imply there’s some universal law that says “licensing OSes makes you rich and not licensing OSes doesn’t”. If I recall correctly, SGI tried doing pretty much everything analysts said it (and Apple) should do over the years, including building Windows boxes, licensing its OS and its software, porting its stuff to Windows, attempting to lower its prices to compete with its cheaper competitors. When SGI released cheaper computers to compete against Apple and Microsoft in the professional graphics market, it ended up cannibalizing its own high-end 3d market and having its lunch eaten by PC workstation vendors and Apple in the “low end”. To the extent that Sun tried similar things, it got similar results. (Solaris didn’t even do well as a free product.)
In fact, it’s hard to find an example of OS licensing working particularly well for anyone except Microsoft, and then only for DOS and Windows (WinCE in its various incarnations has hardly been a resounding success, and XBox is proprietary).
Perhaps the question analysts should ask is why did licensing work (that one time) for Microsoft when it has failed for pretty much everyone else? The answer is pretty simple: Microsoft had a virtual monopoly on office software (Microsoft Word in particular) and refused to port this software to potential rivals, thus preventing them from being viable competitors. (The Amiga, for example, never had a single functional word-processor. I remember standing around in computer stores while small business owners came in and inquired about buying Amigas — which at the time cost half as much as comparable Macs and PCs and had more horsepower — and finally gave up when they saw the word-processor options.) Indeed, I would argue that the only non-MS platform Microsoft supported — Mac — got continued support for a strategically clever reason: alone among rival platforms, the Mac might easily have survived Microsoft’s withdrawal (for a start, it was — and still is — positively bristling with excellent word-processors), and the Mac platform had a long history of providing a profitable incubator for software which — when ported to DOS or Windows — became incredibly popular. E.g. Word, Excel, Powerpoint, Photoshop, Director, and Flash were all “incubated” on the Mac.
Today, Microsoft continues to support the Mac platform — in a desultory fashion — because they can still make money from it, and it’s quite obvious that the Mac platform could cheerfully survive Microsoft’s abandonment of the platform. But imagine if a Mac spreadsheet or word-processor became dominant on the Mac platform — how long would it be before it started challenging MS Office on its own turf?
How Microsoft “Won”
Way back in the late 70s, Microsoft was a small shop that produced BASIC and other languages for a motley assortment of personal computers, including the Apple II and the TRS-80. Among software companies targeting personal computers, it was a player, but that’s about it.
One day IBM decided to enter the personal computer market. It was in a hurry because Apple was in danger of making computers look like serious business tools thanks to a program called Visicalc. IBM didn’t want Apple to have too much time to establish itself since IBM had proven that once entrenched, a vendor with generally inferior products could hold on to a market for a very long time. In its rush to market, it built a computer out of off-the-shelf hardware and off-the-shelf software.
Because IBM licensed MS-DOS rather than buying it outright, and because Microsoft acquired MS-DOS from a third party via dubious means, and because that third party had illegally cloned CP/M (then the most popular operating system for personal computers), IBM ended up with anything but exclusive rights to the operating system. And because IBM computers were nothing special (indeed, Compaq quickly established itself as a quality leader over IBM) and very expensive, it was easy for clone makers to enter the market and for some to become well established, and because Microsoft was better at updating DOS than IBM, and better at marketing than Digital Research (whose DOS was technically superior), Microsoft became a software powerhouse.
Microsoft cemented its position by gradually taking application niches away from rival software companies. It slowly but surely killed off Lotus 1-2-3, Wordperfect, dBase 3, Paradox, Netscape, and Borland’s developer products, bought Powerpoint, and incorporated the features of MORE into Word. It also did its level best to take over other niches, such as desktop publishing and personal finance software, but failed because it wasn’t very good at developing new products and failed to undermine them (e.g. by deliberately making them incompatible with new versions of DOS or Windows) or buy them out.
Perhaps the funniest (in retrospect) thing that helped Microsoft “win” was OS/2. Having been utterly screwed by Microsoft for something like ten years, IBM decided to collaborate with Microsoft on a next generation operating system while Microsoft continued to work on rival products. Of course, at more-or-less the same time, IBM was working on its own version of UNIX (AIX), OpenDoc, and Taligent.
Conclusion
So — based on history and common sense — in order to self destruct, Apple should license Mac OS X, ignore game developers, reduce its prices, release Windows boxes that are cheaper than Mac OS X boxes, port all its application software to Windows, get into a joint venture with Microsoft to release a next generation operating system — or perhaps license Windows 7 and integrate it into Mac OS X, and maybe burn a few billion developing a new OS or two.
Amazingly enough, this list is pretty much a perfect match for what “industry analysts” are telling Apple to do. (OK, no-one is proposing Apple develop a new OS.)
Meanwhile, Apple is hitting record sales and profits, maintaining its prices and margins, and — for once — actively courting game developers and helping them. If Commodore had revved the Amiga every year while cutting prices and had a multitude of top-notch word-processors available, we’d probably all be using Amigas right now. The idea that licensing your operating system is the path to success is simply ridiculous. It worked once for one company in one situation a long time ago.