endless compression

[markw at mohawksoft.com (markw at mohawksoft.com)]

markw at mohawksoft.com wrote:
>
> Not the point, if you have a repeatable compressor, then compression ratio
> is largely unimportant. If you can compress something 3% over and over
> again, then you can make it really small.

>Yes, and if you have a perpetual motion machine, then Middle Eastern oil
> reserves are largely unimportant.

"Perpetual motion" machines are interesting for this debate as we "know"
that they are impossible, however, as advances are made in mechanics and
lubrications, machines are far more efficient. Maybe not perpetual, but
certainly run far longer on the same amount of power.

If something runs at 99.99 percent efficiency, it is not a "perpetual
motion" machine, but it may seem that way to a casual observer.


>Assume that you have a "repeatable compressor" that will losslessly
>shrink any string of bits by 3% every time it is run.  What happens when
>it receives a 100-bit input?  There are 2^100 possible inputs, but only
>2^97 possible outputs.  Therefore, there must be *some* inputs for which
>this compressor cannot produce a 97-bit output--indeed, there must be
>some inputs for which this compressor cannot produce a 99-bit output.
>Therefore, this repeatable compressor cannot exist.  QED.

I have never said there is no end. In each post to you I've said that the
end is when the compressed data is no smaller than the original data.