Ubuntu on Windows

Even more surprised that no one seems to recall that they did exactly this back in the day. There was a version of the NT4 kernel that could run Unix programs. "Great, the program selection of Unix with the stability of windows!" people sneered, but still there was a market for it.

They even marketed NT's POSIX compliance as a great feature.

It was still a little different. It isn't just about building the tools and making them run. It is the environment that comes with it so you can follow any number of recipes you find online to get your development environment up and running. For all the support work I've done in the past, or finding ways to virtualized Linux in a Windows environment (or vice versa), this will be a welcome addition.

But that was a whac-a-mole, probably for getting boxes checked for whatever it is that audit departments did before sarbanes oxley. That they've gone with linux here shows that their focus is mainstream developers. Consider - it their main focus here had been EEE then BSD would have been a far better starting point. The GPL is far more restrictive than the BSD license. The one major area where linux wins over BSD is in drivers and that's irrelevant here. Apart from drivers and package management, BSD is a better platform. They could have found (or bought) a partner in the NetBSD community to do the Canonical stuff.

Now some here would respond to this as - well - what they've done just shows how big the conspiracy is and how far it extends! There's never a full comeback to that. But I think there's a more straightforward explanation.

The straightforward explanation would be this. Windows is no longer relevant in the way it once was - the cloud is the new platform, and people use tablets and phones. And - a lot of developers hate Windows. OSX has emerged as the dominant developer desktop.

They've realised some combination of this, and formed a company directive, "make developers love us".

Even if it means giving developers non-Windows platforms to work with. Like dot net for linux. Or MSSQL for linux. Or free Visual Studio for linux. Or linux that runs on a Windows kernel. Making a solid effort to do linux for azure. And basically giving away Windows as a desktop. Whatever. The things developers need, they're trying to do that. They want to be where the developer are. They're putting their back into it.

The business case is the kind of vague thing that startups take - find some users, make them love us, and then we'll work out how to make money from it.

Imagine the first reaction in the Redmond office when someone read out the feedback post asking for vi and apt-get. Groans all around. But someone in that room responded to the laughter with an "I know, I know" smile and asked, "OK, but what would it take?" And then everyone perked up, and had some fun with the conversation. And they came up with this. And someone saying "shit, we /could/ do that, and it would be amazing!" The people who were in that room will remember that as a career event.

I think it's cool. I want to play with it.

> Consider - it their main focus here had been EEE then BSD would have been a far better starting point. The GPL is far more restrictive than the BSD license.

What Microsoft has created is binary translation for Linux system calls, like Wine but allowing Linux to run on Windows instead. FreeBSD, among many other OSs, has done something similar for a long time, precisely because Linux is the most popular Unix-like OS. In addition, Microsoft doesn't care about the drivers -- everyone still writes drivers for Windows.

Hey, I waved away the existing unix solution, and your link corrects me. Your link shows it has a history far richer than I said: real people put real engineering work into it, often through a fog of unappreciative politics, and now their neglected thing is on its way to being killed. I've been there - those poor bastards. At least a couple of people downvoted you. But it was a worthwhile correction. Thanks.