Ask HN: GWT/PlayN or Unity for a new Facebook/G+ game?

I am making a game with a team of 5 other engineers as a school project. We would like to use either Unity Pro 3 - Education (http://unity3d.com/) or GWT/PlayN (http://code.google.com/p/playn/). We are making a isometric 2D game. It will have a front end on a social network, and a back end for persistence and multiplayer (probably Google App Engine).

Advantages of Unity: it looks like it includes a lot of neat features, including a beefy rendering pipeline and a powerful set of world building tools. (But would we actually use any of that for a 2D game?) It also lets us program in C#, which our team is more comfortable in than Java. It looks like it would make networking super easy (http://unity3d.com/unity/engine/networking).

I don't think our game will be too computationally intensive on the graphics side (no physics), but since it's in a plugin Unity could have fewer performance issues than JS running in the browser?

Unfortunately, Unity has to run as a plugin in the browser, which we're worried will reduce adoption. Also, no one on our team has really used Unity before.

Advantages of GWT: It compiles to HTML5/JS, which is "the new hotness" instead of the "old shittiness" of browser plugins. We have one person on the team with GWT experience, as opposed to no one with Unity experience. Also, GWT-RPC works really nicely with App Engine (most of the time), whereas in Unity I'm not sure how much support we'd have.

However, GWT uses Java, which our team is less happy with than C#. Also, in my experience working with it, GWT can sometimes get heinously obscure error messages, and the only way to fix it is to revert to an earlier version and then try your changes again (in smaller increments).

What do you think?