Re: compression - insights into infinite

On May 31, 6:25 am, jules Gilbert <jules.sto...@xxxxxxxxx> wrote:
This could get interesting...

Look, I have a process, which I check using a tool I had nothing to
with building, a program written many years ago be some very respected
statisticians and communication theorists, and it says that I have an
error rate of 80%. It's used in many labs, all across America and in
Europe.

For a purpose much contrary to lossless compression.

The experiment involves my sending a 4-bit signal across a channel.
One fifth of the time a program correctly guesses the answer.

If one was using that system to compress an e-mail that instructs a
delivery driver where to deliver the items, and the delivery guy has
to guess which one from five possible street addresses is correct,
then that system is worthless as ***, don't ya think? It took the
driver extra hours and fuel to decompress the right address, that's so
not cool.

I am trying to assemble this system into a complete program, and I am
pretty sure that I'm actually saving one bit for each byte I have to
"send" (ie., predict.)

Eh? I'd assume this is one from many systems you attempted. What about
the one in 1996? You trashed it or what when you finally compiled it
and realized it didn't work?

And even if I did, I'd still have to solve my problem with the
patenting system we all have to contend with. 35 years ago I used to
read Computer World and applaud IBM for defending it's patents and
going after bad guys; Well, that law has turned into a morass for us
all, unless you have a thousand lawyers to sic on a company.

Pretty much this is what Microsoft did; After they had stolen what
they were calling "double space" or something similar, and got sued
they put a thousand lawyers on the case -- and minimized their
damages.

Anyway, we compression researchers need more protection from the much
abused laws; (Notice what I am saying: it's the laws that are screwed
up. We need protection from crooks and the law is supposed to protect
us but instead the courts (effectively,) pass laws that enable the
crooks.) This didn't happen because the courts are bad or by design;
It's just been the result of many years of patent litigation and I
have no idea how to fix things.

But the effect is to create an environment that even if I liked and
trusted everyone of you (and I don't!!,) I'd still have to fight off
companies, -- not such as MS, but most certainly MS -- they seem to be
interested in tying up everyone on the planet, their just looking for
the right judge.

Correct me if I'm wrong but you were a neocon last time I checked and
to the best of my knowledge: conservatives favor the capitalist
ideaology that redneck country (US) was founded on. What happened to
your faith and patriotism, Jules?

Libelous litigation and all those predatory tactics you described that
Microsoft and others utilize is the reality of a capitalistic society.
Everyone for themselves. Get fucking used to it.

But in this environment, well -- predicting the future is a whole
easier.

Speaking of which, your stock predictions are not up to par with
something: to predict swings, you'd have to know when sales commence
on the fly. I've been buying two 2L coke bottles a day now for the
past couple months, but a couple days ago I bought 2L of Cplus just to
have something different. You could predict continuing sales of coca-
cola based on the fact that I've been buying for the last couple
months, but HOW would you know I'd still buy a pair tomorrow, or next
year? Are you sure I wouldn't get sick of that nasty caramel-mix ***
a week/month from now? How would you know I'd buy another type of
drink once in a while like Cplus? I never bought CPlus before and
there have been times where I haven't sipped Pepsi for a year.

You can't, 'cuz all these outcomes are *RANDOM.*

But knocking 1-bit off of a byte. Easy enough, and remember my
process doesn't have an extra parameters to carry around, nothing --
no hidden files, no directory entries. Nothing. It's about as raw as
can be.

--jg

That's reassuring.
.