🚀 go-pugleaf

> The only problem I can see with this method is how do you insert data
> in a non-regular pattern and still have the ability to remove it later.

How about:

Send the data in blocks, say 1024 bits.

The first block is 1024 truly random bits, with the constraint that
there is a roughly equal number of one and zeros.  The positions of the
ones in the first block indicates which bits in the second block are
from the text and which are additional random bits.

The random bits from the second block are appended to the first block
and an equal number of bits are deleted from the start of the first
block, leaving 1024 random bits.  These bits indicate which bits in the
third block are text bits and which are random.

etc.

Some possible modifications:

During encoding, if much less than half of the 1024 random bits are ones
at any time, and the new set of random bits doesn't improve the ratio,
complement the new bits.

Use the random bits to do an EXOR of the data.  Depending on how this is done,
it might lead to some correlation in the data.

Insert less than ~50% random bits.

Bob