Why isn’t Backrupr a web service?
It would probably make a lot of sense to have Backrupr as a web service.  However, I believe that the early users would probably like something they can run themselves and manage their own copy of their data, and all of the work that’s been described would be needed to make a good web service anyway.
Why isn’t Backrupr a commercial endeavor?
Two reasons.  First, I’ve got stuff I’m in the middle of at Sun that I’m really excited about, and don’t feel starting a company.  Second, I think this has the highest chance of success as a non-commercial activity, as service providers are most likely to cooperate with it in that form.
Won’t this take a lot of bandwidth for service providers?
It depends on what mechanisms they provide to extract data.  In the best case they’d have an efficient export setup that a good incremental backup scheme could be developed upon.  In the worst case you’d have to screen scape bits of data, which would be massively inefficient.
Will service providers cooperate or try to block Backrupr?
We’ll find out.  One of the reasons for having a rating system is to start to identify who’s cooperating and who’s fighting it.  Hopefully users start to factor that into their decisions and providers will see that they can have efficiency advantages for playing along.
What legal right do users have to the data?
Good question!