i have to agree with Joe. any estimates you make right now will be highly theoretical at best. the xtreme programming (agile development, etc etc) methodology suggests that it is difficult to predict how long a project will take or what, precisely, is involved until you are in the thick of things. the approach i suggest is assume the worst, and get your developer to build a working bare bones prototype as their first order of business. that skeleton app should ALWAYS be functional (via continuous integration using ant or maven plus cruise control or other automation tools).

the payoff for this methodology is that you will always have a functional working product which meets at least some of your requirements right from the start. if there are multiple developers, all of them must obey the Prime Directive: Don't Break The Build!!! this eases integration milestones and decreases the OH-Shit factor.

the more the developers get into developing specific functionality, the closer your estimates will get to reality. then, if you have a deadline approaching, but your project is not "complete", you can approach the application's business sponsor and inform them of what functionality will be ready on any given date. what you have will work, and what you don't have will be a known quantity.