It's just a style preference — I prefer using .sav because I can have both the direct sequence data (raw) and the homopolar sequence data (seq) in a single file.