It takes more than a zillion-dollar marketing budget to build the perfect operating system
Raphael Needleman
I spent August 24 at the beautiful Microsoft campus in Redmond, Washington. The weather was ideal for the launch of Windows 95; the clouds did a perfect imitation of the background graphic on the Windows 95 carton. Five hundred journalists and I gave up our smug attitude for a few hours to sit in the de facto cheering section at an event that reminded many of us more of a fashion show than a computer product launch.
At one point, I had to restrain myself from leaping up in the middle of the hype and screaming, "Listen to me! It's just an OS upgrade! GET A LIFE!" Later, however, I found myself strangely caught up in the act, laughing at Jay Leno's faux-Luddite act and
actually applauding when the big tent wall was dropped to reveal the bleachers full of color-coordinated Windows programmers.
It didn't take me long to come to my senses again, fortunately. All the fancy positioning and expensive marketing can't mask the more important question of the Windows 95 release: How will it really impact the way people work? Because it's the product, not the posture, that will ultimately make the difference.
Will Win 95 take the world by storm, bringing a new class of users to technology? Or will the remnants of DOS that remain in the product push people to more carefully consider the Mac, which has had, as Apple is happy to remind us, a Trashcan icon for 10 years already? Or worse, will the Win 95 hoopla and the emasculated Rolling Stones "Start Me Up" commercial (the words "You make a grown man cry" having been removed from the lyrics) backfire on Microsoft the way the nuns-with-beepers ads submarined OS/2?
I have to admit that I like Win 95. I've been using i
t for several months, and it's pretty good for a major upgrade. But there's a lot of stuff I miss in this OS. I wish I could take the elements I like from all the OSes available and build my own. Component technology doesn't quite offer me that flexibility yet. But if it did, here's what I'd do:
The kernel.
Oh please, let me dump this ancient half-DOS system and build something more tight, more cross-platform, more robust. Like Windows NT.
User interface.
People think I'm a Windows bigot, but that's just because I like to use a lot of different applications, and Windows has more than anybody else. Actually, I'd trade it in for parts of the Mac OS in a heartbeat. The Mac's user interface is better integrated into the OS itself, more consistent, and easier to teach.
Size.
Why do I need 40 MB of hard disk space and 16 MB of RAM for a single-processor, single-user OS? I've seen faster, smaller, and equally graphical OSes that, theoretically
, could do everything I need. GeoWorks and QNX come to mind.
Networkability.
My OS should be able to easily hop onto and off the Internet, a Novell LAN, and wireless services. I like Win 95's capabilities in this arena, but the Mac and OS/2 also do well here.
Multi-everything.
I want flexibility. That means an OS architecture that scales from the address book on my personal digital assistant (PDA) up to a big database server. I want my OS to support multiple processors (which NT and big-iron systems like VMS do). I want it to be multithreaded and multitasking (NT, OS/2, Win 95), so background processes don't bottleneck my whole system. I want it multiuser (Unix), so I can distribute some of my applications without requiring a full-power computer at each seat. Finally, it has to be multiplatform, so I'm not locked into Intel, Motorola, or any other chip architecture.
Object-oriented.
Object-oriented OSes (OOOSes; great acronym, isn't it?)
make for better data sharing among applications and, thanks to reusable components, easier applications development. This last bit is important because it makes programming less expensive, and when you spend less money on that, you can spend more on testing and reliability. The best OOOS we've seen here is NextStep, although the Newton's OS also scores well.
The ideal OS, of course, doesn't exist. But to see how close we are to OS nirvana, check out our Special Report, which begins on page 73.
And in the meantime, be thankful that there is no perfect OS. It means your job as a technologist is safe, as is mine. For now, at least.
Raphael Needleman, Editor in Chief, (
rafe@well.com
)
The Ultimate Operating System?
