Now is the time for inaction. Catch up on unfinished projects. Put out fires. Learn Java. Wait for the emerging technologies to mature and the developing controversies to play out.
Tread cautiously with regard to thin-client strategies based on Citrix WinFrame (a multi-user version of Windows NT 3.51) and Microsoft's Hydra (Windows NT 5.0 merged with what would have been the next version of WinFrame). This technology deserves consideration, but it is already being overhyped and under-scrutinised because Windows is at the centre. Remember, "Microsoft Windows" is like Thorazine to most pundits and research analysts. Basic philosophical concerns about the whole thin-client class of technology vaporise in the tranquillising presence of the words.
If you doubt this, note that many of the criticisms unfairly levied against Network Computers are rightly applicable to thin clients. Yet most of what you read about thin clients contains none of these grievances.
Where are the bold warnings that thin clients are a return to the tyranny of the mainframe? Who is warning you that the IT manager will dictate what software you use? I've seen no predictions that existing networks will crumble under the strain of thin-client computing, in spite of the fact that thin-client computing is more network-intensive than Java-based Network Computing.
I'm not hearing anyone shudder at the risks of having data stored at the server. Where is the whining that thin clients are no more than dead iron if the network goes down? Where is the claim that users will refuse to give up their PCs for these dumb terminals? And why don't we consider it controversial if people contend that thin clients represent lower cost of ownership?
Finally, why aren't people as obsessed with the entry price of thin clients as they were with network computers? The Wyse Winterm 2700SE gets you 486-level Windows performance for about $US1400. Nobody I know has condemned it because you can buy a high-powered Pentium PC for that price.
Somebody switched the price tags
Ironically, thin clients deserve the utmost scrutiny for precisely the same reason they are being excused from it. Microsoft Windows NT is at the centre.
IT managers I correspond with are increasingly disillusioned about Windows NT, and rightly so. Microsoft itself has acknowledged that Windows NT isn't living up to its promises.
That admission comes a little late for some early adopters. One IT manager I know is struggling to cope with the fallout of a corporate decision to rip out all existing Unix and NetWare servers to replace them (along with 10,000 clients) with NT.
After devoting hundreds of people to the task for a year, the company has gained nothing in the way of convenience, cost savings, features, or productivity for its efforts. Rather, the company is now beset with new problems.
The IT staff spends hours looking for work-arounds and downloads because NT lacks an appropriate driver or is missing some of the most basic tools that come free with versions of Unix.
This company is also discovering what many of us have known all along: Windows NT grows increasingly unstable as you install applications. The only cure is to reformat and reinstall Windows NT periodically.
Will Windows NT 5.0 cure some of these problems? I certainly hope so. But guess what you'll have to do to find out? Wait.
But even the shape of Windows NT 5.0 is uncertain. Microsoft is going to be distracted by lawsuits from Sun Microsystems and Caldera. (Indeed, Bill Gates was just recently deposed as part of the Caldera lawsuit to recover damages to DR DOS caused by Microsoft's allegedly anti-competitive behaviour.) And the more I hear about the actions of the US Department of Justice, the less certain I am that its efforts will come to nothing.
Many people threw caution to the wind regarding client/server computing, Windows 95, and Win-dows NT. It's time we learn from our mistakes and begin to tread carefully with eyes wide open.