You have to include an algorithm for decompressing it into enough information to answer a relevant question (in this post's case) "what is the nearest multiple of 10^-n, for some given n?"
Otherwise, it is a (very useful) abstraction. It isn't compression, it is a technique for avoiding expansion in cases where an expanded form is never needed.
Otherwise, it is a (very useful) abstraction. It isn't compression, it is a technique for avoiding expansion in cases where an expanded form is never needed.