Narrowing yourself to 'git users' is a silly thing for two reasons:

1) there are many Windows devs out there who require a VCS 1a) ...who will only use a GUI-based VCS (seriously)

2) there are many users of other VCSes (e.g. Subversion)

These are big markets waiting to be tapped. The answer is not "don't try if nobody else is tapping them". Hacker News is the last place I expect to hear that!

I think it's a massive opportunity. These people are using Subversion because TortoiseSVN is a nice piece of software, not because they are anti-DVCS partisans.

If someone made an app like Tower but for windows and charged $50-100 for corporate licenses, they could make a killing.

Or, if the Github team, having first created a GUI for Mac, which they use themselves and can more easily test, were to then create a Windows version and charge $0 for it, they could convince a lot of Windows Subversion users to switch to Git and Github, paying monthly for private repos.

Which looks like where they're going with this.

github should do some os fingerprinting of tcp connections to its servers and they'd be able to tell fairly accurately what most of their customers use.

That would be very cool.

But I personally think the value in a Windows client would be in bringing new people to github more than satisfying existing users.