r with their own money--cover the complete historical spectrum of computing. Many of these machines are still being used.
Several editors have rather
creaky DOS machines still being used for E-mail (cc:Mail and BIX) and word processing (XyWrite, mostly). Dave Rowell has an unusual Wells American 20-MHz 286. Ed DeJesus has an IBM XT he uses for voice mail (though how he can hear phone messages over the banshee scream of that machine's old hard drive is beyond me). And Tom Halfhill actually travels with a Toshiba T1000SE portable (8086 processor, 1 MB of RAM, 720-KB floppy drive, no hard drive). At home, I have an old Hi-Q 386SX machine (upgraded with a Cyrix 486 CPU) that I use exclusively to support my pathetic attempts to learn to play piano through the Miracle Piano Teaching system.
In our "glass house"--a raised-floor computer room that used to house a VAX but now holds about a dozen microcomputer servers--we run cc:Mail gateways on old reliable DOS computers, and file and print servers on 486s running NetWare. Our Pentiums are mostly reserved for NT and our Web servers.
Why This Matters
This hodgepodge of systems--so
me supported by our IS department, some not--all holds together, through no special architectural design or foresight on the part of software or hardware manufacturers. It holds together because as new technologies come along, we tend to adapt them piece by piece, and fix their integration problems as they occur. And it also holds together because what we need to do here isn't that complex; we're not putting a lot of stress on the capabilities of the systems.
Is this the right way to do things? I would wager that the majority of inexperienced systems administrators would look at our setup and gasp in horror. What is its architecture? Where's the three-year microcomputer plan?
There isn't a very detailed one, and there never can be. Sure, we have a plan; we know what we want to be able to do with our systems in three months, six months, a year. But we'd be kidding ourselves to predict which hardware we will need to support these tasks. ISDN? ATM? Who knows? Who cares?
To really be successf
ul in planning information systems, you have to accept the fact that for a variety of reasons (most of them financial), you're going to be stuck with oddball "legacy" hardware and software for years. It's not a bad thing, either--Ed's old PC voice-mail system is certainly less expensive to operate than a fancy new one. Our cc:Mail gateways work just fine.
Furthermore, beyond some completely obvious areas, you simply cannot predict which specific products you'll need to support in order to realize your broader plans. Netscape or Microsoft Explorer? PowerPC or Pentium Pro? At BYTE, we try to help you answer these tactical questions, but we're always cognizant that these questions wouldn't even be worth asking were we not also asking the more important, larger questions: Electronic commerce or old-style banks? Private or public network services? Constant technology retraining or new hires?
The challenge is to keep one eye on the technology and the other on the bigger trends. You say you want to inv
est in the Web? Are you sure? Perhaps what you really mean is you want to invest in the emerging global electronic community. If so, be sure your plans aren't tied to a particular platform. After all, compared to how hard it is to change the culture of your business to support cyberspace, it's a snap to change technology platforms.
And no matter what hardware you have, it's really hard to learn to play piano.
Raphael Needleman, Editor in Chief, (
rafe@well.com
)
Keep Your Eye Off the Ball
