After he spent 2 years on this project with great success, how come you suddenly know better than him what he should or shouldn't do next?
Breaking into Windows on a solid, non-hacky basis will be crucial to bring NodeJS to the next level. As some other poster said, this will further enlarge the node community and result in even more/better/faster utilities, libs, fixes etc.
Look at it this way, Enterprises are stuck on Windows and their devs don't have a choice about that... but they can influence the development platform that their upcoming projects will run on. Many of those will be highly tempted to break out of the ASP.NET routine. Maybe server-side JS for most is just a shiny toy... maybe they don't "get non-blocking IO" or will resort to old learned threading/concurrency patterns when they shouldn't... but who cares?
Also think about server-side products. FogBugz, HelpSpot, SharePoint ... all provided not just hosted on the cloud but often requested to be installable by admins locally or in a private company-owned/hosted "cloud". Basing future products in those categories on NodeJS will require smooth Windows support to achieve wider acceptance and adoption.
You cannot tell enterprises to change their corporation-wide OS policy... but if you can run your stuff on it, you don't have to embrace the whole rest of the MS stack, they don't care about that one bit.
(IMO, however, the "behind enemy lines" crap is just that. These projects would be a lot better off if they left the poo-flinging to Groklaw and Roy Schestowitz. You want to talk about disheartening...)
(Posted from a Linux laptop while using a Mac as a footrest.)