With the blockbuster technology-sharing agreement between Apple and IBM coming just months after the announcement of the ACE (Advanced Computing Environment) workstation standard, the future landscape of the workstation market is up for grabs, according to industry analysts. The announcements validate the course that NeXT has charted for professional workstations, but they also place NeXT's strategy at risk.
NeXT executives sought to put a positive spin on the developments, pointing out that NeXT already has an object-oriented development environment and a graphical user interface built on top of UNIX. "One of the things we like about the agreement is that it is a real endorsement of our strategy," said Mike Slade, NeXT's executive director of marketing. "They've (Apple and IBM) pre-announced by four years a system that we sell today. It's pretty gratifying."
Indeed, the time frame for the future Apple/IBM operating system is several years out, perhaps as long as five years, said Stewart Alsop, publisher of P.C. Letter. Nor is it definite that any product will result from the preliminary agreement. Several analysts compared the announcement to IBM's aborted plans to build systems around NeXTstep.
IBM licensed the NeXTstep environment in 1988 for a reported $10 million. After initial development on the RS/6000 and the PS/2, IBM abandoned the project. According to Dick Shaeffer, editor of The Technologic Computer Letter, "The parties in the Apple/IBM agreement should take note of what happened with NeXTstep. NeXTstep is a first-class technology, and IBM has done nothing with it."
At the least, Alsop observed, the agreement "clarifies and confirms IBM's lack of interest in NeXTstep."
The 88000 has not fared well in the RISC processor arena, losing market share to Sun's SPARC and MIPS R3000 architectures. MIPS's successor chip, the R4000, has been selected by the ACE consortium, which includes Compaq, Microsoft, and Silicon Graphics, among others. Motorola's last hope for the 88000 was major support from Apple. Shaeffer sees the agreement to let Motorola manufacture the RS/6000 for Apple and IBM as "a way for Motorola to save face" after losing support for the 88000; but in fact, said Shaeffer, the deal "shoves the 88000 into the embedded control market."
Giving Motorola the manufacturing rights for the RS/6000 single-chip implementation is a concession from Apple and a way to maintain its good relationship with Motorola. Apple needs Motorola for its current line of 68000-based computers, Shaeffer said.
Meanwhile, the change in Motorola's direction leaves NeXT in a quandary. It had been widely believed that NeXT planned to follow up its 68000-family computers with an 88000-based workstation. But that would leave NeXT as the only computer manufacturer supporting the chip.
"The 88000 is out of the mainstream. NeXT has enough problems; to come out with an 88000 machine would be bullheaded," said Slater. "To go through the pain of switching architectures and then use the 88000 would be a very foolish move."
However, Shaeffer speculated that NeXT may take a new direction. "Jobs has a knack for headlines," he said. "I think he'll be the first one out with the Intel '586." The Intel chip is the next in line in the 80x86 architecture that has been the mainstay of IBM personal computers and clones. The 80586 is a big step up in power over the '486, however. It uses superscalar RISC technology to produce 100 mips (millions of instructions per second) of performance. Intel has said that the chip will be available in the second half of 1992.
That would be the right time frame for NeXT's next-generation systems. Other chip choices, including the Motorola RS/6000 and MIPS R4000, won't be available soon enough, according to Slater. "The '586 would be a reasonable choice and would give NeXT the ability to run software other than its own," he added.
Whether NeXT now turns to the Intel chip or seeks a different partnership, it is clear that the company has to re-evaluate its existing strategy. NeXT needs to go with a chip architecture that has broad support from other vendors. That's the only way to attract the software developers needed to make the NeXT platform successful.
The good news, analysts agree, is that NeXT has some time. Now that Apple, IBM and ACE have confirmed the wisdom of NeXT's basic approach, NeXT can convincingly argue that it is already delivering technology that the others have only promised. Then it can carefully choose a new chip architecture and still beat its competitors to market with advanced NeXT systems.
Nicholas Baran is formerly West Coast bureau chief for BYTE magazine. Baran is a NeXT user and coeditor of Pen-Based Computing, The Journal of Stylus Systems in Sandpoint, Idaho.