Re: generating function algo

BrainDumb <vortexsny@xxxxxxxxxxx> writes:
Obviously, given random distribution, the only approach is to
represent 1ms with 1bit. So that's 24 * 60 * 60 * 1000 bits

No, the problem is not to encode events or one bit per event, but to encode a given time
with a given resolution.

No, allow me to disagree. The problem is compress data without loss.
My example
is about time sequence, at a millisecond resolution. Details are
insignificant.

I thought somebody knew a clever way to compress set of timestamps.
It appears there isn't a way.

Time stamps are just data. You can't compress data.

Now if you know something special about the distribution of
the data, then perhaps something can be done. However, you've
not mentioned any such special distribution so we have to
presume there's none. I know that in my inertial reference
frame all microseconds are of equal length, and thus it's
equally likely that an arbitrary event will occur in each
microsecond.

Phil
--
Dear aunt, let's set so double the killer delete select all.
-- Microsoft voice recognition live demonstration
.