To paraphrase a famous 17th century playwright: Alas, poor Windows NT, I knew it, Bill: An OS of many blue screens, of most excellent fancy: It hath borne me on its network a million times; and now, how abhorred in our collective memory it is!

Whatever you will say about Windows NT, it was a very successful and prosperous operating system. When Windows 2000 and Active Directory finally replaced it, many NT-administrator types longed for simpler times — as when it was appropriately patched, it was stable, performed well, and was easy to administrate.

But on December 31 the sun set on Windows NT 4.0. Server, with Microsoft ending pay-per-incident and Premier support for all but a handful of well-heeled customers. Microsoft is intent on closing the book on a significant piece of its software history.

But should it?

Almost a year ago, when news of a partial source code leak of the Windows NT and Windows 2000 code base appeared on the Internet, I advocated that Microsoft open source Windows NT. Now, more than ever, I think it should be done.

Never mind the fact that a significant portion of Microsoft's customer base that are perfectly happy with their NT 4.0 boxes are being forced into costly upgrades, now that their support options have been yanked out from under them. An open-source release of NT just plain makes sense.

For starters, open sourcing NT 4.0 would generate some new excitement about the Windows brand, not to mention a tremendous amount of application development from many frustrated Windows developers that are now looking towards Linux and open source as the promised land. With Longhorn and Longhorn Server not likely making an appearance until 2006 and 2007, respectively, that's a long time to go without anything interesting to keep the developer attention.

And unlike Sun's release of Solaris under CDDL, which will be of no interest to most of the Linux folks and will likely generate only a small amount of new development activity, even a quasi-open-source or a restricted open-source license for NT would generate lots of developer excitement.

An "Open NT" could become a very successful platform on its own for open-source software development. Sounds hypocritical? Think of the numbers — few Solaris developers, mucho Windows developers. Do the math. And unlike the case with Solaris, with Open NT, there would be tremendous interest in porting more and more open-source applications to Windows by both the Linux and Windows developer community.

Just for kicks, if I were Microsoft, I'd not only release the NT 4.0 and service pack code base, but I'd also release the Interix software development kit (SDK) under an open-source license, as well.

What's Interix, you say? For those of you who've never downloaded Microsoft's Services for UNIX 3.5, Interix is a whole POSIX-compliant subsystem and development kit for Windows, which allows you to easily port Linux and Unix applications to Windows using open-source development tools. And it really works.

What's in this for Microsoft? Not only would Redmond get a recharged gung-ho developer base interested in porting some really cool stuff, but Microsoft also would generate an entirely new ecosystem of consultants, service providers and system integrators who would have the expertise to build open solutions on Windows.

And since no indemnity- and/or lawsuit-conscious Fortune 500 corporation would ever consider running Open NT, it's not as if Microsoft would lose potential Windows business to a free solution. (If anything, they're going to lose it to RedHat and/or Novell.) I'd just consider open sourcing NT as covering my bases.

Sure, I'm crazy. But I miss my Windows NT.

Jason Perlow is an independent technology consultant and the founder of eGullet.org, the food news and discussion website. He has been integrating (and writing about) Windows, UNIX and Linux networks for over fifteen years.