From a marketing perspective, nsjava is actually a huge win. And the issue isn't the viability of Tcl so much as it is the long-term viability of AOLserver, IMHO. The AOLserver community is just way too small to keep up with the likes of Apache or even Zope in terms of integrating new technologies. It happens to have started with a huge lead on some issues, so it has managed to stay competitive on things like performance and stability, but the new stuff isn't coming on nearly as fast.
What nsjava does, as Dan points out, is it enables AOLserver to leverage work being done in other Open Source communities. And it does it in a way that lets OpenACS developers keep using Tcl. As a bonus, we get to be able to tell folks who have already made big investments in Java apps that AOLserver can play nice with their existing toys.
