Apple marked its 20th anniversary with a major change in direction: the acquisition of NeXT Software. Although Apple highlights five rather abstract points behind its acquisition of NeXT, most of the Macintosh community is concerned with two things: what a Apple-NeXT operating system will do, and when it can be delivered.
First, a caveat: don't expect a definitive evaluation of the Apple-NeXT merger from TidBITS - or any industry publication - for some time. Hindsight may be 20-20, but for the next several months both Apple and the Macintosh community will be staring ahead into a thick blanket of fog. By acquiring NeXT, Apple has defined the arena in which it intends to play, but the first real test of the Apple-NeXT merger will probably be at the World Wide Developers Conference in May of 1997. Until then, all bets are off.
What Happened When -- With the plethora of announcements, punditry, and vacation time in the last two weeks, it can be difficult to grasp how the Apple/NeXT merger came about. Here's a chronology that should clarify matters:
Aug-96: Apple shelves Mac OS 8, codenamed Copland, in favor of biannual incremental system updates that will incorporate some Copland technology. Developers and users clamor for more details, but aside from talking up the incremental updates, Apple is tight-lipped. Be demos a version of the Be OS for Power Macintosh, and rumors of Apple-Be talks emerge.
Sep-96: Rumors of Apple-Be negotiations ignite press speculation about Apple's OS future. NeXT is mentioned in passing.
Nov-96: Speculation about Apple and Be is fueled by mostly-unconfirmed stories in mainstream and trade press. Though Apple and Be struggle with rumor control, Apple's silence on its operating system plans leaves more questions than answers.
Dec-96: Major Macintosh publications like MacUser and Macworld, struggling to say something about Apple's OS plans, run extensive features on the Be OS.
19-Dec-96: Serious rumors begin to circulate that Apple and NeXT have reached an agreement.
<http://live.apple.com/next/ 961220.pr.rel.next.html>
23-Dec-96: Apple and NeXT executives meet with the press to answer questions not covered at the initial announcement. Although some answers were helpful, some were vague or contradictory.
The Plan (As We Know It) -- As outlined prior to the Macworld Expo, Apple's operating system strategy can be summarized as follows:
Apple will continue to develop and enhance System 7.x on schedule for at least the next 18 months, with the first incremental update (Harmony, or Mac OS 7.6) shipping this month. Additional Mac OS 7.x releases are scheduled for July of 1997, and January of next year, although there will undoubtedly be updated components available in between. Although some developers are openly apprehensive, applications for System 7.x have at least two years with a large (probably growing) market of installed users.
These plans don't mean you should start surfing the Web for screenshots of the NeXT OS, thinking you're looking deep into the eyes of a future Mac. No matter what else, Apple must produce a Macintosh. Apple's now-legendary ease-of-use and elegant design must be apparent in any future operating system if Apple hopes to compete with Windows NT and other mainstream systems. That doesn't mean abandoning good ideas from NeXT, but rather making sure the polish and style of the Macintosh are unequivocally present.
The NeXT Hurdles -- Why are the specifics of a Mac-NeXT operating system so vague? Simple: Apple and NeXT haven't had time to evaluate the complex technical issues at the heart of such a project.
First, there's the question of a kernel. A kernel is a relatively small bit of software that coordinates nearly all services for an operating system, including memory allocation, process management, data access, and interaction with hardware. OpenStep, conversely, is a comprehensive software layer designed to run on top of multiple kernels: it presently runs on the Mach kernel (a Unix variant) and Windows NT, and can deploy to other flavors of Unix.
One of Apple's first tasks in designing a Mac-NeXT operating system will be to decide what kernel to use as a building block on Macintosh hardware. Numerous advocates would love to see Unix at the heart of the Macintosh, and note that Apple has been sponsoring development of a version of Linux for Power Macintosh based on the Mach kernel. Since OpenStep is already available for Unix, many people believe Mach is an obvious, best path for Apple.
<http://www.mklinux.apple.com/>
So far as Macintosh hardware is concerned, my personal bet would be on NuKernel, Apple's own microkernel developed for Copland. Although NuKernel (obviously) never shipped, Apple has been working on it for some time, and it was designed with Apple's existing hardware (PowerBooks and various AV capabilities) and software (QuickTime, OpenDoc) in mind. However, OpenStep could provide the possibility of developing a future Mac OS designed for non-Macintosh hardware, and applications written for the Macintosh could conceivably run under other operating systems.
Another fundamental question for Apple regards display technology. NeXTstep relies on Display PostScript to generate its screen display. This has always been one of the "gee-whiz" features of NeXT, and graphics professionals have long wished for Display PostScript on the Macintosh. However, Apple's acquisition of NeXT doesn't mean Display PostScript is necessarily going to become part of the Mac OS. First, Adobe owns Display PostScript, and currently collects royalties for every copy of Display PostScript sold, just as it collects royalties for every PostScript printer sold. Apple might understandably be wary of entering into a licensing agreement for Display PostScript, which would not only put the Mac OS at the mercy of Adobe technology, but would increase the already-high cost of the Mac OS. Second, Apple already owns a graphics and typography engine that is (in many ways) technically superior to Display PostScript: QuickDraw GX. However, developer support for QuickDraw GX has been lackluster (for a variety of reasons). So far, Apple hasn't said anything about the fate of Display PostScript or QuickDraw GX in future operating systems.
Compatibility -- An issue of primary concern is backward compatibility with current applications. Macintosh users will (rightfully) expect the vast majority of current and future Macintosh applications to run on a future Mac-NeXT operating system. Apple would do well to remember that a significant portion of its customer loyalty stems from strong backward compatibility. A Mac Plus can still run System 7.5.5, and many users (myself included) run software written in 1988 or earlier. Although Apple may attempt to bridge rank-and-file customers to a new operating over a two or three year period, at a minimum that new operating system will have to be reasonably compatible with System 7.x applications.
A key to backward compatibility might be the work Apple has already invested in MAE, a Macintosh application layer that essentially lets Mac OS applications run on a variety of Unix workstations. MAE is not faultless, but it does provide a surprising number of Macintosh services on non-Apple hardware. Technical knowledge from Apple's MAE effort may lead to strong backward compatibility for System 7.x applications under a future Mac-NeXT operating system, even on non-Macintosh hardware and/or non-Apple kernels.
Vendors and Technology Support -- It seems likely many major Apple technologies will be available for a future Mac-NeXT system. Some of these - like QuickTime - represent some of Apple's most-coveted properties, and Apple has a tremendous investment in making sure first-rate implementations run on Apple's operating systems. Similarly, Apple is likely to maintain its commitment to OpenDoc, which (in theory) could benefit significantly from a Mac-NeXT system. Less clear are the futures of technologies like Speech Recognition, QuickDraw 3D, and (particularly) QuickDraw GX. Other core Apple technologies - AppleTalk, Apple events - will be essential to making the new system walk and talk like a Macintosh.
However, less certain is how well third-party developers will take to a new operating system. Several development tool companies (including Metrowerks, Symantec, Altura, and Tenon Intersystems) have already announced intentions to support a Mac-NeXT system; no doubt numerous additional commitments will be forthcoming from vendors at Macworld Expo. However, announcing plans to support a forthcoming system is one thing: delivering products is another. Apple and the Macintosh community probably won't have a good idea how solid support for a Mac-NeXT operating system is until after the World Wide Developers Conference in May.
What to Expect at Macworld Expo -- Let's face it: Apple and NeXT haven't had time to make solid decisions about many of the issues surrounding a Mac-NeXT system. As a result, you can expect Apple to speak of schedules in vague terms ("first half of 1998"), and to make general statements about commitments to backward compatibility, core technologies, and key Apple markets. Apple's objectives for this week's Macworld Expo are simple: demonstrate momentum toward a modern operating system; ensure that attendees feel good about the NeXT acquisition; and rally third-party vendor and developer support.
What About Be? If you think Jean-Louis Gassee and Be will cool their heels in the wake of the Apple-NeXT merger, think again: Be plans to move forward aggressively with the Be OS. Be OS DR 8 for Power Macintosh is now available for developers via the CodeWarrior for Be OS package, the January issue of MacTech magazine, and with Power Computing systems beginning this quarter. Furthermore, Be is widely expected to demonstrate some Macintosh applications running under a Mac OS emulator from fredlabs for Be OS on Power Computing machines.
<http://www.be.com/>
<http://www.powercc.com/>
<http://web.xplain.com/mactech.com/>
Many technology writers have expressed disappointment that Apple didn't purchase Be OS as a foundation for future Macintosh systems. I think Be is in an enviable position and may have the better end of the deal: they remain a nimble, forward-thinking company with a genuinely exciting product and strong momentum. I sincerely hope that Be and Apple will continue to work together toward the success of both companies.