Developers at a crossroads: How will Apple's incorporation of Next impact the future of NextStep for Solaris?
Will the merger help or hinder Sun's marketshare?By Rick Cook
The acquisition of Next by Apple might well have given Apple the proverbial shot in the arm it needed: a high-end, high-performance operating system that works. For Sun/Solaris developers and IS managers, this begs the question: Should I be concerned or happy? The answer is: both. (2,400 words)
Few mergers in recent years have caught the imagination of so many as has the union of Apple and Next. However, this merger is important in the Sun world for two significant reasons. First, Next is a major supplier of object-oriented development tools for the Solaris environment, as well as co-developer of the OpenStep object API that ships with Solaris. Second, if Apple succeeds fully (against admittedly long odds) with its incorporation of Next's operating system into its installed base and new installations, it could have a significant impact on Solaris' future.
Since the Christmas announcement of the purchase of Next by Apple, events have come thick and fast. Apple announced its new operating system based on Next and set an ambitious timetable that has the first developers getting copies of the new OS this summer. Next is already history. "I've already got my Apple badge," says John Landwehr, product marketing manager for what used to be Next and is now formally Apple's Redwood City, CA, facility.
The object of Apple's affection
The first general release of the new operating system, called "Rhapsody," will be early next year with the first mass market release coming in the summer of 1998. This is very fast for a major software project. This is, in part, because it has to be if Apple is going to survive.
The polite way of putting it is to say that Apple faces major challenges. The less polite way, as described by The Wall Street Journal, is to compare the position of Apple enthusiasts today to that of Tory loyalists in 1784 -- after Cornwallis surrendered.
A combination of declining marketshare and aging technology has been dragging Apple down for the past several years -- so much so that a year ago the rumor was that Sun would buy out Apple. Perhaps Apple's low point came in August 1996 when Apple CEO Gil Amelio had to announce that Copland, Apple's new operating system, was being dropped. Apple's stock and expectations promptly did the same.
The current Mac OS is based on a 10-year-old architecture and lacks modern features such as pre-emptive multitasking, multithreading, and symmetrical multiprocessor support. Also in the past decade, Microsoft, Sun, and others have increasingly copied the features of the Apple user interface (itself copied from work done at Xerox's Palo Alto Research Center, PARC), reducing some of Apple's differentiation from the rest of the desktop community.
Apple has been trying to come up with a modern, object-oriented operating system for the Macintosh for several years, and Copland was its second failure. Its first was Pink, an operating system developed by Taligent, a joint Apple/IBM venture.
Rhapsody, the new operating system, promises all the features Apple needs. "Rhapsody will be buzzword compliant," says Russell Brady, an Apple platform and technologies spokesman, with his tongue firmly in his cheek. "It will be multitasking, multiprocessing, and all the other cool stuff." In addition, he boasts, Rhapsody will be clearly superior in Internet and multimedia support.
Rhapsody will get all this by being built on Next's Mach kernel and by offering the Next API and Next object-oriented development tools. This is a well-developed package that Next has sold successfully to software developers on multiple platforms for several years.
Developers on platforms such as Solaris are extremely important to Next and represent a significant potential revenue stream for Apple. Since Next stopped trying to sell its own computers, it has become one of the biggest players in object-oriented development tools. Apple and Next took great care to reassure Next developers that the merger would have only beneficial effects on them.
"Apple will maintain Next's commitment to cross-platform and cross-processor support," Apple CEO Gil Amelio promised in an open letter to Next customers that quickly followed the announcement. "We will continue to develop, sell, and support products currently available, including those for Windows NT, Solaris, HP-UX, and NextStep." The letter also said that cross-platform support for WebObjects and OpenStep "aligns perfectly with Apple's overall strategy."
In general, Next has been under-appreciated, in part because its attempt to follow the Macintosh model by being both a hardware and software company failed. This was partially due to the fact that Next's operating system had not gained major marketshare -- and partially because of the antics of Steve Jobs, Next's mercurial founder.
In the development community, the story is quite a bit different. Next is a highly-regarded platform for developing software. It has a powerful, object-oriented set of development tools, as well as one of the best Web suites around in WebObjects. The Next operating system is built on the Mach microkernel and includes a lot of up-to-date Unix features, such as threads.
Taking advantage of the microkernel
Although Next has a complete operating system, most of the software isn't written to run on the Next operating system. In true microkernel fashion, Next does almost everything through its NextStep API. Next ports NextStep to new operating systems, and software written to it runs on the new OS with a recompile.
The Next architecture is also portable and has a reputation for being quite productive. It runs on Solaris, NT, and other operating systems with a very high degree of API compatibility, so much so that Next claims developers can maintain a single development tree for NextStep applications. Next enthusiasts claim that Next 's object-oriented nature and development tools significantly reduce time to market for projects developed with Next.
Among those enthusiasts is Tim Berners-Lee, the father of the World Wide Web. Lee points out that the original Web software was developed with Next. "At CERN in 1990, I developed the first client and server World Wide Web app using Next's tools in just two months," he says in a quote displayed on Next 's Web site. "It was a combination of Next's object-oriented environment, neat development tools, and a powerful class library which made it all possible."
One of the major worries for Next developers is that they will get lost in the focus on Rhapsody development. This is especially true for the companies that are not writing for the PowerPC processor, the basis of Apple's modern Macintosh line. This includes the developers who use Next's system to write for Solaris on SPARC. These are the people Apple and Next are working hard to reassure.
For Next developers, the upside of the merger is that if Rhapsody works, they will have instant access to a much larger market for their products. Apple may have less than 10 percent of the U.S. PC market, but that is still a user base in the millions. This is much larger than developers of other environments (such as Solaris) are used to, and all those users will be available to Next developers almost automatically.
There's a downside to the upside as well -- two of them, in fact. The first is that just as Rhapsody opens the Macintosh market to Next developers, it also opens their markets, such as Solaris, to Macintosh developers writing for Rhapsody. That includes software heavyweights such as Claris, Adobe, and Symantec which might take advantage of Rhapsody to offer their wares on platforms such as SPARC.
The second part of the downside is that the intercompatibility simply may not mean much for either Next or Macintosh developers because the markets are so different. To the extent that the Mac is a general-purpose personal computer (as distinct from its presence in specialty markets such as, say, graphics) the applications tend to be much more mass-market oriented. They are less expensive but also generally less capable and simpler. Except for a few highly developed personal productivity applications, such as word processors and spreadsheets, Mac and Next/Unix software buyers might well stay separate, minimizing the impact of Rhapsody.
What's Next for Sun?
For Sun, there's also the question of what will happen if Rhapsody lives up to Apple's hopes. If Rhapsody meets Apple's expectations for technology and market impact, the influence on Solaris could be significant. As proposed, Rhapsody hits nearly square in the center of Solaris territory. It would be a powerful, modern, Unix-derived operating system with strong object-oriented development tools and Web presence. Potentially this could cut into Solaris' market.
For that to happen, Rhapsody has to succeed both technologically and in the market. This is where the long odds come in.
Oddly, the technology seems to be less of a worry. Although Rhapsody is quite advanced and a major project, it is not being built from scratch. "We've done a lot of ports of NextStep and OpenStep in the past," Next's Landwehr points out. "We had a lot of work done for the PowerPC port because we were going to ship our own Next PowerPC back when we were in the hardware business." Further, the Next development team is generally highly regarded, and they have a lot of experience porting Next to other platforms. And of course, well-managed object-oriented software development projects are well known for their speed -- especially if the team is experienced.
The worrying issue technologically is that traditionally it requires a couple of years to work the bugs out of the first release of a new operating system. A modern operating system is very complex, and it's not easy to get it right. Microsoft, for example, has been taken several years to get a solid version of its Windows 95, which is why a lot of IS managers are opting for the more stable Windows NT. Sun's experience with Solaris has been similar.
Microsoft can afford to use its customers as beta testers. Apple can't. Rhapsody must be significantly bug-free in its first wide-scale release at the end of this year, or it will lose precious momentum and probably end up like OS/2. Apple is confident it can do that. A lot of people are skeptical.
Most of that skepticism is based on experience with Microsoft. Rhapsody will be object-oriented with all that implies for fast, clean software development. Still, it is a major technical hurdle and one Apple must clear cleanly.
Optimists like to point out that Avi Tevanian, the former Next vice president of engineering who is now in charge of developing Rhapsody, has a reputation for getting working products out the door. (They also like to point out that Steve Jobs will not be directly involved with developing Rhapsody.)
Still, even Apple acknowledges it won't be easy. "There are concerns," Brady admits. "Those are valid concerns because we didn't deliver on all we were going to do with Copland. We know the industry's watching us. We know we'd better deliver, and we know our track record has been patchy in the past."
Market acceptance of Rhapsody is more problematic. Essentially Apple is attempting to recreate its success with the original Macintosh by using clearly superior software on non-standard hardware to attract customers. One of the questions is, can the strategy work again in the late '90s under quite different market conditions?
One good sign is Apple is avoiding the mistake IBM made with OS/2 by not tying the new operating system to existing hardware. This gives Apple more freedom to exploit the new software and avoids the swamp of backwards compatibility questions and performance compromises that plagued the first versions of OS/2. Once Apple gets the first release of Rhapsody out the door, it will go back and add compatibility features for existing Macintosh systems for the first "full release," slated for next summer.
Is there an upside for Sun?
If Rhapsody is even two-thirds of what Apple expects, it will be a very good operating system. The question is, will it be good enough? Will Rhapsody be so superior that general computer users will overlook the incompatibility with the existing Wintel platforms and switch to Rhapsody in large numbers?
On the surface this is also Sun's problem with Solaris, especially Solaris on SPARC. However, the dynamics are quite different because Apple has to sell a lot more computers to make its model work. Unlike Sun, which has always been a "workstation company" with relatively low volumes at relatively high prices, Apple is generally a mass-market company, and it needs to move more units to succeed.
Technically, Microsoft is certainly vulnerable. Windows 95 and NT are not object oriented -- or much of anything else except widely used. Under the covers they are both complex and do things in odd ways. As one industry watcher describes OLE, "This isn't an API, this is a shaggy dog story in C!" But technology isn't the determining factor in operating selection, or everyone would be using something like Solaris.
For Sun and Solaris, a successful-but-not-overwhelming Apple Rhapsody would be an overall advantage. It would boost advanced operating systems and give Microsoft something else to worry about without taking marketshare from Solaris.
Apple's success is a long way from a sure thing, but in evaluating what you hear in the next year there are a couple of things to keep in mind.
First, Cassandras to the contrary, Apple is not on its last legs. The company is in long-term trouble, and it has blundered badly in the past two years, especially in the matter of a new operating system, but it is still a $10 billion company with a 26 million installed base. It has a core of loyal followers and a solid hold on such niche markets as graphics and publishing. In other words, it is more akin to Digital Equipment Corp. than Wang just now.
Second, Apple has pulled this off before. In the early '80s when the IBM PC and MS-DOS were just taking hold, Apple rebounded from the disappointing Apple III and the expensive Lisa with the Macintosh, sparking a software revolution and assuring Apple's position as a major player in the computer business. What Apple needs to do is repeat its success of the early '80s in the mid-90s.
Remember, too, that "success" for Apple doesn't necessarily mean market dominance. The Rhapsody Macintosh doesn't have to become the best-selling computer to be hugely successful and meet Apple's goals. At the very least, Apple needs something that will help it maintain its position. Ideally, it should be able to grab additional marketshare as well.
About the author
Rick Cook is a regular columnist in our sister publication, NetscapeWorld. He divides his time between writing about the Web, computers and high technology, and novels. He can be reached at firstname.lastname@example.org