Interested in an alternative engine in dejadup?

I've put together a backup engine, analogous to rsync --link-dest, in Python. It's called Backshift.

I've been looking for a backup scheduler system/GUI that uses rsync, that might be interested in adding support for backshift as an alternative to rsync.

Is Dejadup perhaps such a project?

Here's a URL to the project:
http://stromberg.dnsalias.org/~dstromberg/backshift/

The next feature I'm planning to add, is the ability to do backups without messing up a system's buffer cache - since almost by definition, backups read a large number of files once, and then don't read them again anytime soon.

Right now, backshift has:
1) Hybrid fullsaves/incrementals, like rsync --link-dest
2) The ability to chop files up into variable-length, content-based blocks; this is more effective than fixed-width blocking, though slower
3) The ability to deduplicate those blocks without requiring untoward amounts of VM
4) The ability to compress those blocks with xz (via ctypes, subprocess or os.popen) or bzip2
5) The ability to resume interrupted backups
6) The ability to expire old blocks and old metadata

It runs on CPython 2.x, CPython 3.x, PyPy and Jython, is thoroughly documented, is thoroughly pylint'd, and has a comprehensive automated test suite. It's also been tested on a very large number of operating systems, including many *ix's, OS/X and Windows 7 with Cygwin.

Files are selected using find (I've been considering adding a tar-like way of saying "Do everything in this directory, recursively" though), and restored by automatically assembling a tar archive from the compressed blocks, on the fly. This maximizes the use of other popular tools like find and tar, and allows one to restore via the network using just ssh and tar.

?