Special Report: Mac OS 8 - Back to Square One

Apple has killed the Copland OS as we knew it. What's in store now?

By Galen Gruman and Anita Epler

November 1996

Copland, we hardly knew ye. For two years, Apple has painted a brave new world for the Mac OS, one that would bring a state-of-the-art infrastructure to today's OS, delivering agents to help automate your work, preemptive multitasking and threading to let programs do several things simultaneously, and an interface that would both simplify and customize your interactions with the Macintosh.


As Macworld revealed exclusively a month ago, Apple has dismantled its grand OS scheme. CEO Gilbert Amelio confirmed Apple's plans at the Boston Macworld Expo a few weeks ago, and AppleSoft vice-president Jim Gable has shared with Macworld more details on the new OS strategy.

The new strategy means that users will face a constant stream of OS updates: Apple plans to release new versions of system software quarterly, as well as semiannual upgrades with substantial new technologies, interspersed with smaller bug-fix releases.

The new strategy also means that Copland–officially called Mac OS 8–and its follow-on, Gershwin, are essentially gone. Many of the technologies originally scheduled to be released with Copland and Gershwin will be developed for new versions of the Mac OS, but there is no grand target for what the future Mac OS will become. Instead, Apple's OS development will be more evolutionary, based on the technologies Apple can deliver when it can deliver them. Apple's new strategy is a mixed blessing–it's bad that there's no grand target to motivate users, developers, or Apple; it's good that there won't be broken promises.

The Copland name will still be used, but only to refer to groupings of technologies that Apple plans to deliver–Copland no longer means Mac OS 8, just the technologies once promised as Mac OS 8.

The current strategy change stems from decisions by Apple's new chief technology officer and vice-president of research and development, Ellen Hancock, who decided to end the mega-release Mac OS 8 strategy and focus on delivering real enhancements in a more timely way. That decision led to the incremental-update plan, supported by Amelio, and to the reorganization of engineers from several Mac OS teams (Harmony, Copland, and so forth) into one Apple Mac OS group.

Apple is being coy about its plans for the coming years, but Macworld has gleaned the following from interviews with Apple employees and from internal planning memos. Keep in mind that Apple's official line is often out of sync with its unofficial line, and that Apple is revisiting most of its strategies, so the company may again change its OS approach.

Some changes may come from outside pressure. The latest set of announcements has angered outside developers who are now unsure what to aim for; inside Apple, numerous engineers and product managers are also unclear on what the future holds, and many have seen their pet projects delayed or canceled. Because of these reactions, Apple may reconsider its changes.

System 7.5 Shell Game

According to Apple's original schedule, we'd all be running Mac OS 8 today. But months ago, Apple realized it needed to do something in the interim, so it launched Harmony, a System 7.X version scheduled for early 1997, which was originally slated to incorporate parts of the Copland interface and wrap in revised versions of QuickTime, OpenDoc, Cyberdog, QuickDraw 3D, and the Finder. Apple officials presented detailed Harmony plans this May at the Worldwide Developers Conference.

That was May. In August, Harmony was clearly behind schedule, so Apple broke it into two pieces: one for January 1997 and one for summer 1997. The January System 7.X piece (perhaps to be called 7.6) will include updated versions of QuickTime, OpenDoc, Cyberdog, QuickDraw 3D, and the virtual memory manager, but no Copland technologies. "We'll have a few things there to wrap [these components] all together," says AppleSoft's Jim Gable, "but essentially there'll be no new stuff in January."

So why bother with the January release? "This is a big deal for Live Objects developers," said Gable. (Live Objects is the new name for OpenDoc components; the engine and the architecture retain the name OpenDoc.) "With this release they can expect the infrastructure to be in place that they need," without having to include OpenDoc with their products. Developers won't need a license from Apple to distribute OpenDoc.

The System 7.Y release–slated for July 1997–will be more in line with previous reports of Harmony's features. "This is where we'll start to see Copland technologies," says Gable. Apple still isn't sure what Copland features it will roll out in July. Gable states, "We've got to do some homework to find out what we can include. We don't know how much we can decouple." This is frustrating for users and developers alike, but probably the right thing to do considering Apple's poor track record in predicting OS components.

Mac OS 8's Shadowy Future

So when will Mac OS 8 ship? No one knows. The official date is late 1997, but Gable acknowledges that this is optimistic: "I don't know when it will ship."

When does the name change to Mac OS 8? No one knows that, either. Gable says it's no sure bet when the OS number will change from 7 to 8. The logical time to change the number is when the new microkernel is incorporated into the Mac OS, since the microkernel is the guts of the OS, and changing it is what gives Apple the ability to add most of the Copland features that so excited us a year ago: preemptive multitasking, protected memory, and so forth.

But Gable says no decision has been made on what will constitute the number change. It could be the microkernel, or it may simply be when Apple runs out of decimal places for System 7.X.

Copland as originally defined is gone, and so is Gershwin, the planned successor to Copland. Copland was supposed to be a way station to Gershwin, with Copland being the Mac OS that introduced preemptive multitasking, multithreading, and protected memory–all key to seamlessly operating simultaneous programs and to reducing system crashes and application conflicts. Copland was supposed to be like Windows 95, in which these services are available for the operating system but applications still share memory and have to multitask cooperatively (as they do in System 7). Gershwin was to introduce preemptive multitasking, multithreading, and protected memory between applications, so each would reside in its own safe space (as in OS/2 and Windows NT). The good news is that by not shipping Copland/Gershwin as originally planned, the Mac OS might see full preemptive multitasking earlier than the original Gershwin schedule.

Because Mac OS 8 remains undefined, no one–developers or Apple engineers–really knows what to aim for. The effect has been dramatic. Finally poised for the first developer (beta) release of Copland–known as DR1–in early August, Apple pulled the plug unexpectedly during Macworld Expo. "We're not going to release it as it is," Gable says, while confirming that those developers who work on low-level code will receive some subset of the current beta and that engineering work will continue on Copland DR1.

Gable says rank-and-file applications developers will be seeded for interim OS releases but won't receive DR1. "Because we're not going to be doing a single, monolithic release, there's no reason to release what we have," he says. "There's a tremendous amount of work in there, [but] a single, monolithic seed is wrong."

Some developers have interpreted the DR1 pullback as a sign that Apple's commitment to them has waned. But Apple's Gil Amelio acknowledges that the company will have to work closely with developers to ensure that they have the tools they need to support the system updates.

The OS That Keeps On Coming

In an interview with Macworld, Apple's CTO Hancock made it clear that Apple can't keep giving away the Mac OS. Apple needs to find a way to make money from it to pay for its continued development, she asserts. This subscription-like release plan may be one of those ways.

Gable notes that one advantage of the new incremental-release plan is that many Copland technologies will now run on 68040-based Macs, 680X0 Macs with PowerPC upgrade cards, and perhaps even some 030-based Macs. (Under the original plan, Copland wouldn't run on any of these.) Now that there is no unified mega-release, Apple can decide on a component-by-component basis whether a technology will perform on 680X0 Macs.

"Now you don't have to wait for the next major release to get neat new stuff," Amelio says. He points to the Mac OS's complexity as justification for rolling out the next operating system piece by piece.

Of course, there's danger for Apple, even if it has no choice: major architectural changes–such as incorporating a new microkernel or adding preemptive multitasking–generally don't lend themselves to being rolled out in piecemeal fashion. "That's one of the things we have to figure out," admits Gable. "No one really knows for sure."

The Last Word

The Mac OS is now officially a moving target. By necessity, Apple has converted Mac OS 8 from a goal to a journey. It's not at all clear where that journey will take us. That's unsettling. Apple needs to illuminate the path it's on, without again setting us up with unrealistic expectations. Right now, we have no expectations. In the coming months, let's hope that Apple gives us a new set of Mac OS expectations so we know the platform will continue to move forward, without inflating our hopes and then dashing them.

Senior associate editor ANITA EPLER and executive editor GALEN GRUMAN both follow Mac systems developments closely. Additional reporting by Howard Baldwin.

Copyright 1996