You might want to consider offering a couple of different options; none of these file transfer services have a FB requirement:
http://wetransfer.com http://yousendit.com http://letscrate.com http://ge.tt
Or perhaps you want to position this as a FB-only service? It would certainly differentiate your service, but might also limit your user base unnecessarily.
Technical question: how are they able to do this? I needed to programmatically send Facebook PMs a while back, and from what I could tell it was not possible. What gives?
Limiting user base? I don't think it would hurt much. Half of internet users use Facebook. I think they have a good chance at giving better user experience by tying with Facebook.
Even assuming that figure is true, why cut your potential user base in half right off the bat? And for those who are FB users, what do they do when they want to exchange files with non-FB users? The service will be unusable in such cases even for FB users, and sendrail ends up losing more transactions.
I know that if a third party service wants to tie to Facebook, I immediately say "No thanks".
Same here. I tend to assume that any company so coupled to Facebook probably has the same concern about user privacy as Facebook, i.e. none.
If half of those internet users friends use facebook, this limits Sendrail to 1/4 of potential matches.
If half of those internet users' potential recipients of files are friends (colleagues, acquaintances, chat room contacts, etc may not be fb friends) that gets us to 1/8 of use cases as potentially solved with Sendrail, and Sendrail is only useful to Fb users 1/4 of the time.
Most people won't want to use a service that's useful only 1/4 of the time for a particular task, so this is far worse than a 50% reduction in potential users. This is how you get no users.