* Refactor web.fetch.simple to use per connection timeouts via the timeout kwarg for mechanize.open

* Refactor IPC code to use communication logic from multiprocessing

* Use multiprocessing for cpu_count instead of QThread

* Rationalize books table. Add a pubdate column, remove the uri column (and associated support in add_books) and  convert series_index to a float.

* Replace single application stuff with Listener from multiprocessing

* Refactor add books to use a separate process named calibre-worker-add
  - Dont use the process for adding a single book
  - Use a process pool for speed

* Change mobi metadata setter to use author_sort setting from MOBI output plugin instead of mobi.py