LBFS:
Q1: LBFS decides where boundaries lie based on the actual data contents. Why does it 
match the hash of the sliding window of contents to a fixed predetermined value?
i.e. Why not just directly match the sliding window content to a fixed value?

Q2: If I open a bitmap file, insert a small object somewhere in the picture and
save it under a different file name.  Will the original and newly saved files
have much in common according to LBFS? How about text files?

Shark:
Q1: Why does Shark name contents with cryptographic hashes instead of using i-node # or NFS file handles?