Dojo (HowTo)







  Easter Eggs




  Martial Arts

The truth about how software development really works

By:David K. Every
©Copyright 1999

The Press loves to point out that Copland (Apples Mac OS 8 project) is a failure. However, they seem not to understand the realities of software engineering, and do themselves a disservice each time they propagate this myth.

Copland as a monolithic project has died, but Copland as far as a collection of technologies, designs and source code is certainly not dead. Many have already shipped, and many more are scheduled to ship in Mac OS 8, and many more later on in Allegro (Mac OS 9) or in Rhapsody (Apple's next system).

I realize that it is too much to hope for to expect reporters to stick with the facts and explain what has really happened - so that is why it falls to us non-reporters.

What was Copland and where did it go?

Apple started scheduling a big new monolithic OS release (code named Copland) around '93/'94, with another follow-on OS (Gershwin). Copland was a transitional OS meant to add new technologies (while keeping compatibility with System 7) on the way to the fully protected, preemptive, multiprocessing OS called Gershwin. The target was for Apple and Copland to beat Microsoft to market with their Windows95 project -- this lasted about a month, then a reasonable goal was set and the target date of "end of '96" (with Beta's going to developers in May of '95). Gershwin was scheduled for '98 (loosely).

If you take Apple's target date "End of '96", and compare to when MacOS 8 did ship "Mid '97" (with the rest shipping Mid '98), Apple was not very far off at all -- in fact Apple was far closer than Microsofts first projections for Chicago / Win95 (set for late '92, but shipped in mid '95) and much better than Cairo (scheduled for '93, but still not shipping, nor scheduled to any time soon). Apple did drop features from Copland for Mac OS 8 -- but they also added many many more. Microsoft dropped many features from Windows 95 and added none.

The press over looks all these facts and instead sells it as "Microsoft is a success -- Apple is a failure." Yet the truth is that --

  • Microsoft sold under half of their projected sales, Apple sold over double theirs.
  • Microsoft shipped 3 years late and dropped many features, Apple shipped 6 - 18 months late and added features (as well as dropping some).
  • Microsoft has met much resistance (to this day) -- and there are still far more Win3.1 machines than Win95 machines. Apple is already getting pretty good acceptance for Mac OS 8 -- and with far fewer user install problems and far more market acceptance.
  • Apples features actually work, while many of Microsofts (like PnP, 32 Bit, or preemptive multi-tasking) do not always work.

Apple announced to developers what Copland was, and what technologies would be incorporated and showed the pieces that they had and sold developers on pieces that they would have. But then things got messy. Apple licensed the Mac, the internet started taking off, Multi-Processing became much more important, RAM prices fell through the floor, and people started screaming less about compatibility and footprint - and cried instead for more for power.

So the goals of Copland changed, and then changed again. The scheduling got changed and things got mixed up. But Apple did ship many (most) of the Copland technologies, or is about to. Other technologies got added, and other just changed (when and how). Here is a partial list of some of those technologies -

  • Apple decided that with licensing such a big issue that the HAL and PPCP support (scheduled for Gershwin) was made the highest priority and moved to Copland (this was later changed to System 7.6 and Mac OS 8 - with parts of it shipping before that).
  • Copland was scheduled to have some OpenDoc integration, and for Gershwin to have really tight OpenDoc integration. 7.6 already has more OpenDoc integration than was promised in Copland.
  • Internet integration was added (7.6) this is CyberDog - shipped, but then OpenDoc got changed because Java took off. Apple still got that back in with Mac OS 8.
  • OpenTransport - shipped
  • OpenTransport/PPP - shipped
  • QuickDraw3D was scheduled for Copland - shipped
  • QuickDraw3D RAVE was a new addition - shipped
  • PCI Drivers - shipped
  • DR-Recompiler - shipped
  • updates to the Memory Manager - shipped
  • updates Virtual Memory Model - shipped
  • V-Twin (Apples cool search technology) - this has been added to CyberDog and will be added to Mac OS 8.
  • New Native Finder (shipped with Mac OS 8).
  • New Filing System (HFS+) -- shipping with Mac OS 8.01.
  • Making most of these technologies work on 68K machines - new - so far many have shipped and were never promised.

So what hasn't shipped?

  • NuKernel

Now a Kernel is a big part of an Operating System, but it is still just a part. And yes, that part got delayed for an extra year (scheduled for 1998), but they are still going to try to get in in there with Allegro (Mac OS 9?). (There is still a question as to whether they are going to use the NuKernel or Mach Kernel, but they are planning to add it).

Copland had changed so much that it was not going to be recognizable as the original, and many parts of Copland had been taken out and already shipped. Apple had decided to continue to break out the rest and ship each technology as soon as parts were ready - instead of holding back important technologies for some big mondo-hype release like Win95. Apple also decided to make most of these completed technologies work on both PowerPC's and 68K machines (Apple had only promised PPC support for these technologies). At this point it was silly to keep the Copland name - when the only thing that was going to be left was the new kernel, and Apple had changed directions to bypass a Copland like release and instead to jump right to the Gershwin design (and then some) with Rhapsody -- which is STILL going to ship in 1998.

So the names and priorities changed - and some code and work was lost - but most of the work WILL be leveraged. The goals as a products (Copland/Gershwin) were wrong - but much of the engineering effort will be repackaged under a new name (like Rhapsody). This is not a failure, but responding to the demands of the marketplace and of Apple's developers (1) .

(1) I was one of those developers screaming for this direction -- so I am glad they changed.


It is easier for the press to refer to Copland as a failure, rather than explain the realities. However, when a company; leverages a lot of code and design from a project, and uses that to get me more features and functions sooner (rather than later) and delivers functions on more platforms than was originally promised, and makes such dramatic changes (in response to my needs) -- I call that a success.

Everything I wanted out of Copland is shipping (or will be) -- some parts shipped early, others late, and many new functions were added. Gershwin (now Rhapsody) is still going to achieve its original target ship date, and it is going to deliver FAR more than Apple promised. So I am very happy, and see the success of Apple's OS strategy --no matter how misinformed the press and public is, or how bad the spin gets.

Created: 02/09/97
Updated: 11/09/02

Top of page

Top of Section