>>> I'd imagine few users are still running NT4 workstations; they are typically power users who would have upgraded to W2K or XP long ago. <<
>
>That might be what you imagine, but in fact there are lots of companies still running NT 4. Contrary to what Microsoft would have us believe, many companies don't run right out and buy the latest version any more. They have come to realize that the cost of upgrading is significant when training time and learning curves are factored in. And the track record for stability in new OS versions is less than stellar.

First, let's be clear about what we mean when we say "running NT 4". I agree completely that lots are still running NT4 servers - I have one client still running on NT 3.51. However, half of my business is installing and administering networks, and not one workstation at any of my clients is still running NT4 Workstation.

Also in my experience, Windows 2000 is more stable than Windows NT, based on server installations. XP is not yet available in server form, and none of my users run it 24/7 so I can't really comment on its stability; however others on this forum have found it better than W2K Pro.

Your point about upgrade and training costs of a new OS is well taken. However, I'd argue that the purpose of the OS is to support applications. If a company is contemplating an app upgrade that would also require an OS upgrade, factor in part of that OS upgrade cost when determining ROI.

It's likely to be mid- to late 2003 before VFP8 SP1 is out and/or developers start deploying significant apps. Ask yourself, how many of your clients will still be running NT4WS and not contemplating an upgrade?

Ask yourself if it's worth forking your development efforts for these few, likely hidebound, customers.

If they're your main customers, ask yourself if you shouldn't remain in VFP7, which would avoid forking and give them the time of your next upgrade cycle to upgrade their OS.