People want to know when the project will be done.  That's reasonable.  But attempting to list all the necessary tasks and estimate the amount of time for each one is doomed to failure.

Estimate in relative terms.  I don't even recommend using "ideal engineering days."  Pick one small story and estimate others relative to that.

Other estimation approaches:
 * [[http://www.mountaingoatsoftware.com/products/planning-poker|Planning Poker]]
 * [[http://agileworks.blogspot.com/2008/01/team-estimation-game-by-steve-bockman.html|Team Estimation Game]]

Then see how much you get done in an iteration (or other time period).  Chart this over time and visually estimate when you'll be done.  It can be that simple.