When you play bridge, you sometimes overbid. It might be a sacrifice, to prevent your opponents from scoring. Maybe the cards just look good despite the point count. Probably, it’s your partner’s fault, but you’re stuck playing the hand.

Sometimes, you’ll make the contract anyway, which might make you greedy. Then you’ll overbid the next several dozen hands, too.

Your opponents will love you for this. Your partner, however, just might become your ex-partner.

Microsoft isn’t in the bridge-playing business. It has, however, overbid its hand, and it might just be time for you to become its ex-partner.

It is, in short, time to seriously consider Linux as a desktop operating system. If it isn’t time to develop a migration plan, it’s certainly time to perform a cost/benefit analysis.

It’s a triple irony: The timing coincides with Dell dropping its support for Linux; it comes just as Microsoft has finally achieved sufficient OS stability to eliminate that as a differentiator, and it is Microsoft, through its licensing shenanigans — not any of Linux’s backers — that has made Linux a credible option for you.

Microsoft has backed off from its upgrade-by-October-or-else marketing plan for Office XP. So what? Its willingness to play licensing hardball with customers is beyond doubt. Its preference for renting software rather than selling it is overt. And the correlation between its products’ marketshare and pricing is a matter of record.

Which means you no longer know what you’re going to pay next year for Microsoft Office. Think you can evade the problem by avoiding upgrades? Think again: Whether through business growth or retiring older machines with OEM licenses, you’ll need new copies of MS Office, and you’ll buy them under Microsoft’s terms.

The problem is, you don’t know today what Microsoft’s terms will be tomorrow.

Linux lets you avoid these risks. That doesn’t make the migration free and easy. IT and end-user training are obvious costs. The soft costs of reduced productivity and time spent tweaking imperfectly converted files are others.

The hard part, though, will be remembering that despite rumors to the contrary, personal computers are important because they’re personal. You’ll need to help end-users convert or find Linux correspondents to those nasty unsupported or personally developed applications so many IT managers try to prevent in the Windows environment.

If you don’t, an unsupported personal information manager will be the least of your problems. You’ll see renegade copies of Windows appearing on desktops throughout your company.

Now that would be ironic.