While I agree that migrating data for users is something that would be of value, having anaconda do it is not really a great idea. I view migrating data as a different task than installation. Knowledge of where settings are stored and what data is actually important or what the user cares about is very application and OS dependent. It's not something we should do during installation, but rather offer it as something the user can do after installation. They may find that Fedora is not even something they want to continue using.
We already set up dual boot systems in Fedora just fine. Creating a new tool that can handle data migration would be helpful for users, but definitely out of scope of the installation process. When I start thinking about migrating data from another operating system, I think about these questions:
- Should each application know how to import data from other, similar programs?
- Should there be a central system that knows how to migrate data between operating systems or environments?
- Really, aside from files such as documents and music (i.e., files that can be easily moved between platforms without much hassle), what things are we talking about migrating?
Zoltanh721, good idea but I'd like to see implemented outside of anaconda. Oh, and if a project like this takes off, don't rely on partition IDs to determine what other operating systems are on the system. They mean nothing. libblkid will be more helpful.