Re: compression - insights into infinite

On May 31, 11:51 am, Jim Leonard <MobyGa...@xxxxxxxxx> wrote:
On May 31, 7:25 am, jules Gilbert <jules.sto...@xxxxxxxxx> wrote:

The actual data is a byte, another process (a duplicate of the 80%-
failure process,)  reduces the value to a residue, ie., numbers like 4
or 5, plus a sign bit.  Pretty good I think.  Then I use the
information associated with this process to resend the right-most
portion of the byte, and find that I can predict it at the receiver
side about 20% of the time.

There is nothing wrong with this system; it's just that you've
neglected to realize that, as you approach entropy, keeping track of
the "residue" eats up any space you may have saved through prediction.

Thanks, Jim. But I have read the relevant posts, going back years,
and I have decided that I do need to be a little more persuasive. By
providing some confirming insights.

I have developed a fundamentally new class of compressor (which is
after all, a very interesting type of transform, mathematically.) But
I need to convince others that I actually have done this, and that
means demonstations and the like. Not showing the code, but at least
one or two private demos.

Recently I was invited to demonstrate the program to a US government
organization -- I really do need to "get out there" or at least send
my nominees.

I've been intending to make additional information (examples,)
available but I'm just so busy with a bunch of other stuff. But I do
intend to do this. In fact I spent the day preparing some files.

And I really could use some help with that book, "Probabilistic Metric
Spaces" -- I know that a few of you are excellent mathematicians, all
I am looking for is someone who can help me through two chapters. So
come on...

--jg

.