Node.js beginner questions

I have been really playing with node.js alot here lately. I really like it and would like to get serious about building a site that I have in mind with it. There are just a few things that I had questions on. I am currently using express, cluster, mongoose(mongodb), and dust because it seems like the community is doing pretty well with this setup.

The first question that I have run into is making changes in the app. If I change one of the files in my app (with the exception of templates), I have to restart the app in order for the changes to show up. Is there something out there that will do this automatically or is there something I am missing?

The second question I have is, are there any really good resources out there that anyone can recommend. Seems like alot of the stuff that I have found through google just explains one topic and there are not alot of books on it yet(see some good ones releasing this year though). I was wondering if there was a really good resource for node applications (really interested in how people are structuring there applications and what they are using).

The third question or really feedback I would like to get is what good node modules should I be looking into using for my application. I really like cluster and running apache benchmarks with it versus without it I saw a double in the number of requests it could handle when using all 4 cores versus just the single core. I want to try and keep my code so that I dont have to write the logic twice for the backend / frontend which is why i have decided to use dust as my templating choice.

The last thing that I would like to know is based on an article that I read by linkedin (http://engineering.linkedin.com/frontend/leaving-jsps-dust-moving-linkedin-dustjs-client-side-templates). After reading this article I was just impressed by the ingenuity that they came up with to do this. I see many benefits from this by saving bandwidth on the server from not having to server process the files each time and letting the users browser do the load. It also will allow for less bandwidth throughput by not sending the whole template to the browser every time and caching it. But I am just stumped as to how they accomplished this. Can someone please shine some light on this subject or if you know of a good article that explains how to do this it would be great. I dont know for sure I would use this implementation but I would love to know how to do in case it is applicable.

Thanks for any feedback.

First question: Actualy there are multiple tools that do the job starting from nodemon to supervisor

Second question: Learning materials

Third question: i think my Second question answer includes this one also.