Here's what I've experienced.
I download a pilot episode of a series which I happen to like (after watching the episode).
Then I download a nzb which has the full first season (including the pilot)
When I remove the pilot from the download job (defined by importing the nzb) it says its incomplete (duh I deleted the pilot) but it doesn't post process all the other episodes, even if they're verified

So now I still have to manually extract them and remove all the left-over files.
(also when retrying the failed download in this example, it rechecks EVERY episode even though they're already been checked, but this should be solved using the below method)
This is what I'd like to see:
Do not define post processing jobs by imported nzb's, define them by par2 collections (download jobs can still be defined by the imported nzb).
In my opinion there are two advantages when doing it like this:
1. It post-processes each individual episode when the others are still being downloaded
2. It post-processes each individual episode when others might fail.
I know I could just import each episode in a separate nzb and have no problem, but sometimes this isn't always available or preferred.
With 'job' I mean a row in the queue or history tables, and with 'define' I mean creating a job.
Greets,
n2k