2002 - 1997 = 5

Apple had a ready and rock solid OS to release in 1997 but they didn't release it. Instead they started tearing it apart only to build it up again and release it five years later. What did Apple do in those years - the five most crucial market years in the history of personal computing - that could ever justify not releasing their OS sooner?

Get It Try It

It's not a particularly well guarded secret that Avie Tevanian, head of software at NeXT and Apple, voted against Apple releasing Rhapsody, voting instead for creating an entirely new operating system.

OS X was created instead - not a new system but a mixture of Rhapsody and new and old ideas thrown back in.

Why couldn't Apple release Rhapsody in 1997 - and beat Microsoft to the punch? Beat Microsoft's Windows 98, Windows 2000, Windows Me, and Windows XP? What was so important Apple sacrificed the chance for major market share for five full years?

Apple fans like to think they're always ahead of Microsoft but it's most often the other way around. When it comes to secure 32-bit computing Redmond beat Cupertino by a full ten years. But Apple had a chance to halve that down to five years in 1997 - before Microsoft had full control of the market - and yet they didn't take it.

The ink dried on the Apple NeXT 'merger' in December 1996 and by January 1997 300 NeXT engineers - including Jon Rubinstein and Avie Tevanian - had made their way from Redwood City to Cupertino. By 1997 OpenStep was a force to be reckoned with. Steve Jobs himself walked in the door with OpenStep on an Intel laptop. OpenStep was the only OS capable of giving Bill Gates a run for his money.

OpenStep/Rhapsody was running on everyone's hardware. On Sun's Solaris, on Hewlett Packard's HP-UX, on its own FreeBSD - and even on Windows NT. It was a given winner. And NeXT had finally turned things around and were looking at steady profits for the first time in their corporate history.

And Steve Jobs had more or less prophesied the outcome years earlier when he speculated that NeXT would either be the last company capable of maintaining a hardware lock-in or the first company to fail at it. And in failing - and changing their name from NeXT Computer to NeXT Software - they'd finally found their niche.

They had the new gorilla of the personal computer OS market.

Everybody wanted OpenStep - and for good reason. IBM had been a scratch mark away from shipping it instead of OS/2. They'd also invested heavily in its development as had many corporations. And the investments were finally paying off.

And then NeXT packed up and moved to Cupertino. But the operating system survived the grueling journey and was still market ready. Why wait another five years?

Barring an intelligent answer to that question: what did Apple do for all of those five years? Those five years in which Bill Gates and Microsoft consolidated their stranglehold on the personal computer market for the foreseeable future?

Those five years in which Microsoft released a total of five major updates to their two branches of the Windows operating system - Windows 98, Windows 98SE, Windows Me, Windows 2000, and Windows XP. In those five years Apple did nothing.

Nothing in the OS marketplace at least: they let Microsoft win by default. They gave the game away to Microsoft.

Was this part of some big strategy so secret to this day even the media pundits can't figure it out? Was there a clever tactic in letting one's already precarious market share slip even more? Is there a way no corporate strategist yet knows about to win back an entire market by first losing it completely?

What exactly were Apple doing between 1997 and 2002? More precisely: what were they doing with OS X?

And what were Apple's shareholders doing?

Apple did release new products in the period 1997 - 2002 but nothing on the scale of the system they already had in house. They released updates to the irrelevant and increasingly wobbly MacOS and the first iMac which sold a few million units. But the new OS everybody needed was nowhere to be seen.

Microsoft went to secure 32-bit computing in 1992 with 'Windows NT'. They had interfaces to all the big Unix database server applications. NeXT had them as well. But Apple were still crashing with their cooperative multitasking system and would still be crashing five years later. And yet now they had OpenStep and the 'rock solid foundation of Unix' - something Microsoft couldn't beat. Didn't Apple want the market back? Didn't they want the profits?

Didn't their shareholders want them?

So what were Apple doing with OpenStep - what later would be called OS X?

• The Apple HFS file system had to be adapted to OpenStep. This was in practice an impossible task but they made a go of it. In particular OpenStep (Unix) file system APIs had to be wired into HFS which had to be ready with Unix answers to Unix questions.
• The Unix hard link feature was a particular quandary that was never solved. The partial solution they arrived at means that today OS X has a secret directory in its file system that almost no one can get at or see.
• The extremely user friendly cascading menus used in OpenStep were replaced with the more familiar but far inferior Apple Mac menu bar.
• The extremely sophisticated OpenStep dock (replete with tabs) was replaced with something simple - some would say 'too simple'.
• OpenStep's submenus were detachable: they could be 'ripped off' and moved anywhere on screen. This code embedded in the OpenStep frameworks had to be ferreted out and removed.
• The OpenStep document controller - the spider running the object oriented OS X desktop - had to be rewired to understand Apple's HFS file system.
• The OpenStep frameworks themselves had to be rewired to allow for 'toll free bridging' for legacy Apple Mac applications that wanted to run on the new platform. This meant taking code comfortably embedded in the object oriented frameworks and creating procedural APIs from it - code the frameworks would be calling in the future. These new APIs had to be designed in such a way even legacy code could grapple with them.
• In effect where OpenStep had a single API OS X was to have two: the one from OpenStep (albeit surgically altered) and a new API for non-native code.
• Adobe had come out with an 'update' to EPS called PDF which was an even better choice for vector graphics. Apple rewired the OS to use PDF instead.
• A new look was necessary from a commercial standpoint. This look would be called 'Aqua' and it would be extremely successful.
• Scroll bars had to be moved from the left to the right. Voices who insisted this be a configurable user setting were drowned out by Steve Jobs who - after years of using scroll bars on the left with OpenStep - suddenly proclaimed that scroll bars must be on the right and only on the right.
• Drawers and sheets were invented - two cool new UI window types.
• A somewhat new application architecture had to be invented to accommodate legacy Apple Mac code.
• Some Unix command line tools had to be modified to accommodate Apple's legacy HFS file system.
• The developer tools had to be augmented to accommodate Apple's legacy HFS file system.
• Project Builder had to be modified to accommodate legacy Apple Mac projects.
• Interface Builder had to be modified to accommodate legacy Apple Mac windows and controls.
• Transition tools had to be created to help legacy ISVs take the step up.
• The OpenStep file manager had to be gutted because no one at Apple liked it. It was dropped in favour of an update to Apple's old Mac Finder which essentially made it identical with OpenStep's file manager.

And so on. Many of these changes were inevitable and would have occurred no matter where the work was being done, be it Redwood City or Cupertino. Many of the changes were however not inevitable, could only occur in Cupertino, and are the sole reason the OS could not be released until 2002 - despite its being more than ready in 1997.

Of course this glosses over thousands upon thousands of small changes made throughout the system - changes that would have been unnecessary and even undesirable if Apple had simply taken OpenStep/Rhapsody anno 1997 and thrown it out the door.

Another way of finding the answer to the unanswerable question would be to ask Apple or anyone totally in love with Apple or Steve Jobs exactly what Apple added to the mix between 1997 and 2002 that NeXT wouldn't have added a lot sooner. Follow up that question by asking what Apple removed from the mix.

For OS X users, OS X developers, and Apple shareholders those are important questions to ask.