I meant to say in my last e-mail that I understand what you were trying
to do. I have also tried this extensively for the past year (which is
how I learned about data compression and the various approaches).
I have tried storing things like the max, min, mean, median , standard
deviation, etc. of the data. Basically what are known as various order
statistics. What I found out through a lot of painful trial and error
was that the original data could not be reconstructed UNIQUELY from
this data, unless the statistics that I was storing was equal to or
greater than the original data in terms of data size. In other words,
to store 64 bits of data (say eight 8-bit integers), I needed to store
more than 64-bits worth of statistics (usually a lot more), and that
didn't always guarantee that I could reconstruct the data uniquely!
A lot of work has been done in the data compression field by a lot of
really smart people. I would trust that they have done a good job and
implement Huffman coding or a variant that doesn't require a license,
unless of course your project would benefit from the advantages of
Arithmetic coding and you have a budget for the license...
Ken Fleisher
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
