Issues

Outer-space medley

In which the editor unburdens himself on Plan 9, on NT (again!), and on Unix magazines.

One of the more charmingly quaint sights at Uniforum last March was the AT&T booth, where samples of the experimental new Unix, Plan 9, were being given away. I looked forward to loading the system on my PC at home and checking it out, but alas, the founders of Unix are still entirely focused on technology and oblivious to the needs of the uninformed public. I was able to load Plan 9 onto my hard drive, but the system could not recognize my Microsoft mouse; the man pages weren't where the instruction sheet said they were; the instructions for booting it from DOS were incorrect; the files copied to the PC hard drive are invisible so I can't find them; and there is no tech support, of course. (One of my editors couldn't even get this far, and he's been a Unix nerd all his life.) So, on paper Plan 9 may be an insight into the future of operating systems, but on my hard drive it's still from outer space.

No universal Unix
At he heart of my December editorial "The End of NT" was the observation that the server market is fundamentally different from the mass desktop market. I've discovered an added argument supporting my case: NT won't be 80 percent of the enterprise server market because the server needs of enterprises are too varied to be incorporated in any one server system. In other words, it's impossible for there to be one universal Unix that serves all needs.

That being true, the Unix vendors need to market their OS, first, by emphasizing conformance to standards at specific core levels, as in SPEC 1170, and second, by strongly marketing the performance and feature and experience benefits of each given Unix against each other Unix and NT. For example, Sun can talk about its long experience in SMP; Digital Equipment can claim clustering expertise; IBM can brag about its discardable kernel and parallel-processing experience; HP can talk about its mainframe-mimicking features. No one vendor will dominate, because no one factor dominates user needs, and no one vendor can afford to evolve its Unix in all directions at once. Instead, different Unix implementations will emphasize expertise, tools, and capability of greatest value to different customer subsets. This strategy revives the argument that in its (properly conceived) diversity lies Unix's greatest strength. And that realization becomes Unix's greatest defense against the illusion of the unstoppable NT.

Trouble in Paradise
Open Systems Today folds into Information Week, Client/Server Today folds into Datamation -- is the Unix market so unhealthy? No, to the contrary, the Unix market is in the best shape of its 25-year history. But as Unix becomes more mainstream, vendor advertising grows in the mainstream sector. Vendors still think they can overlook the Unix professionals -- but they're wrong. The vendors who understand the importance of the Unix professional will prosper over those who don't.

OST's marketeers argue that the IS market is becoming a homogeneous, undifferentiated whole. This is muddled thinking. The islands-of-computing model is indeed rapidly becoming obsolete, but that's not the same as saying computer professionals are now all the same. There will always be professional and technical differentiation in the face of complex user needs and complex technologies to meet those needs. Buying-decision power will remain in the hands of those who manage and understand the technology. And I believe there will always be a need for publications like Advanced Systems that provide a view of the world of computing from the perspective of each of its core competencies.

About the author
Michael McCarthy (michael.mccarthy@advanced.com) is editor in chief of Advanced Systems.

If you have problems with this magazine, contact webmaster@sunworld.com
URL: http://www.sunworld.com/asm-05-1995/asm-05-issues.html.
Last updated: 1 May 1995.