Copland, Where Are Thee?

By Don Crabb

Dateline: May, 1994. Apple promises developers at WWDC 94 that the next generation operating system, Copland, will be in their hands as DR1 by year’s end for testing and released to their customers by mid-1995.

Dateline: May, 1995. Apple promises developers at WWDC 95 that the next generation operating system, Copland, will be in their hands as DR1 for testing by the fall of 1995 and released to their customers by mid-1996.

Dateline: October, 1995. Apple figures it’s high time to tell us all the truth: they don’t know when Copland will ship, save to say it surely won’t ship in 1996 and might ship in early 1997.

With this kind of product development cycle and advance planning it would be safer for all of us if we simply read the entrails of slaughtered farm animals to predict the availability of Copland. Maybe we could bring the Oracle at Delphi out of retirement? Or blow a call in to Jean Dixon?

My point, of course, is that our friends in Cupertino have done it again! They have mismanaged a development project that only has the fate of the company and its loyal developers’ lives hanging in the balance. If it weren’t so serious, it would be hysterically funny.

I don’t imagine, however, that any of you are laughing. I know that I am not. In fact, I imagine that you’re probably even more pissed than I am. Not because Apple is behind in its Copland development efforts, per se, but because they have deliberately misled us and our customers.

The Truth Hurts

Apple knew months ago that they wouldn’t make any of the 1995 and 1996 Copland dates. They knew they were very late in Finder and microkernel development. They knew they were even later in the development of other core technologies, including OpenDoc, Open Transport, and PowerTalk 2.0. They knew all this and they didn’t bother to utter a peep. Nary a nod in our direction. And that, folks, is wrong.

It’s wrong (as well as stupid) for Apple to mislead its developers, much less its customers. It’s wrong for Apple to mislead its tool developers. And it’s wrong for Apple to mislead itself - a problem clearly evident when you start talking with any of the PowerMac engineering teams. Many of these folks were floored when they heard the news about Copland slipping yet again, since they were recently told things were right on schedule. It’s bad enough when Apple lies to us, but it’s bonehead city when it lies to itself.

What We Know

This means, of course, that you will not be getting your DR1 release this fall as you were told in May. If you work at Symantec, Metrowerks, or one of the other tool vendors, you will get something that Apple’s calling the Copland Tools Development Release, a DR0 CD with very early versions of the Finder and microkernel on it, as well as Copland APIs.

DR0 might be enough for Greg Dow, Greg Galanos, Jean Belanger, and the ultrasmart braintrust at Metrowerks (hell, it will have to be!) to produce the CodeWarrior for Copland release, but it’s nowhere near enough along for us to do anything with it seriously (besides, we’d still have to wait for CodeWarrior, anyway).

All this means we might get our DR1 in March of 1996, but that release will only interest those of us doing device drivers and some extensions. DR2, which will have more general utility for the bulk of Mac developers, won’t grace our CD-ROM drives until at least May of 1996 (and that’s probably another overly positive scenario - some engineers on the Copland team are saying DR2 won’t hit until October of 1996).

Speculating Onward

This puts a DR3 or more robust DR2 into our hands maybe by the end of 1996, but more likely early 1997. This release might also be made available to large Mac sites for testing. Finally, by mid-1997, we’d have the released Copland 1.0 - only two years later than were told we’d get it at WWDC 94, and a year off the schedule given us at WWDC 95. And all of this speculation assumes no further serious slippage, which we all know just isn’t the case when it comes to very complex operating system development.

What Might Have Been

All of which makes me wonder why the hell Apple didn’t collapse some of this development cycle by buying or borrowing some of its basic services for Copland. Certainly the experience of moving from System 6 to System 7 must have taught them something. And the move to Copland from System 7.5 is a couple of orders of magnitude more complex than that transition was, making it all the more important to get a boost on the process.

I keep wondering if we all wouldn’t be well on our way to System 9 or 10 by now if Apple had only used the Pink (Taligent) kernel from IBM as their point of MacOS departure, rather than building the Copland microkernel from the ground up.

Hoping for the Best

Of course, all of this is water under the bridge now. And we are still stuck in the same position we’ve always been: hoping that Apple is telling us the truth and being honest to itself about its own development cycle.

Given that Copland requires the integration of complex existing and developing Apple OS technologies (QuickDraw GX, PowerTalk, OpenDoc, just to name three) into a code base that’s still not an alpha, the betting money is on Copland’s release date to customers slipping even more than the latest Apple prognostications would allow.

That puts the release of Copland interestingly close to the release of the next version of Windows NT (Cairo), and likely several update releases into Windows 95 (and/or the Windows 95 release that migrates everyone to the new NT). And it pushes Gershwin out until perhaps the turn of the century.

I wonder if by then any Apple customers will care that Copland shipped sometime in 1997?