On the one hand, I have been writing my package (directly into the .m file) separately using the standard Mathematica front-end. My source .m files and the init.m files have the appropriate folder structure to make the package work by simply running
<< myPackage` It all works very nicely.
On the other hand, I am building the complete documentation of this package with Wolfram Workbench 2.0. The intention is that when I'm finally done, somehow I could manually take all the documentation files generated by Workbench, and put it into the same folder where I've been developing my actual package. Then hopefully the documentation would be properly integrated into the documentation center, and I can distribute this folder to others. However, after writing up the documentation for a handful of functions (which apparently is a ton of work), I realized how flimsy Workbench is, and I am beginning to feel that most likely I am not going to be able to figure it out in the end because there appears to be so much going behind the scenes every time I "Build" the documentation [and there is also a "Deploy Application" button which I'm scared of pressing].
What are the steps to integrate the documentation generated by Wolfram Workbench with a pre-existing package? In particular, which files do I need to move into the package directory? Do I need to be careful about the name given to the project in Workbench? What if the various functions in the package have been defined in different sub-pacakges (different sub-contexts)? Would it still work?