[UFO Chicago] CEO's Report for January 24th, 2002

[John Kilbourne]

I'd say that you can't talk about compression without talking about 
extraction, and something is compressed if it has to be extracted. 
The perlscript is never really extracted, unless the code has to be 
maintained. Then the spaces have to be put back in. The 
encycyclopedia is not extracted, unless it becomes so hard to read 
without margins that some other operation needs to be performed to 
make it readable. In other words, there is an "original format" and 
a "compressed format". 

My two cents.

PS Thanks for the dinner Eliott.

John


------------------ Reply Separator --------------------
Originally From: "Peter A. Peterson II" <pedro@tastytronic.net>
Subject: [UFO Chicago] CEO's Report for January 24th, 2002
Date: 01/25/2002 12:24pm


CEO's Report for January 24th, 2002
-----------------------------------

Well, last night's meeting was a success -- we had

Nate Riffe, Jack Edwards, Elliot Shank, Jordan Bettis, Mike McCune,
Jack Beglinger, Neil Ormos, Ian Bicking, John Kilbourne...

and ME!

That's 10! Plus, a small cameo by none other than Ron Sweeny, his
wife, and his son, Quentin.

We ordered 3 pizzas: Large pan cheese, large pan sausage, and one
EXTRA LARGE JOSE SPECIAL! (The waitress heard "Pedro Special" and
misremembered it as "Jose Special".)

We looked at some TRON extras on the DVD and had a demonstration of
Nate Riffe's cool USB storage device.

But the real meat of the meeting didn't happen until later on. 

I'm not sure exactly how it started, but Jack was talking about laser
discs, and the difference between CAV and CLV laserdiscs (Constant
Angular Velocity and Constant Linear Velocity), two different formats
of Laserdiscs. I can't remember which is which, but the idea is that
on one format, each "ring" of the laserdisc is a different frame,
which allows for great freeze frames, but stinks for space
considerations, because that means that the ring closest to the
spindle has the same amount of data (well, it's analog...) as the ring
furthest from the spindle. This means tht that that as you approach
the outer edge of the LD, the data becomes much less dense. The other
format said, hey, we can increase the space on these things if we
double up, and put 3 frames on the outer edge, where the innermost
edge only has one frame! Since this is analog, you could compare this
to stretching or "compressing" the signal to fit into a different
space.

Neil Ormos called this a compressed format, and then all hell broke
loose.

So the new mission of UFO Chicago is to define what is, and what is
not compression. We couldn't do it last night, and if none of us had
real lives, and didn't need sleep, we'd probably still be haggling
about it right now. Needless to say, I suggest we take this discussion
to the list until the next meeting.

In order to facilitate that, I shall present some of the basics of the
conversation, and we can fill in, or pick up, from there.

There were several analogies presented:

* "If you cut the margins off of an encyclopedia, this is a form of
compression." This is called "The Encyclopedia Margin Analogy": If you
are taking a Laserdisc and making the data take less space in order ot
conserve space on the disc, you are compressing the data. This is
analogous to cutting the margins off of an encyclopedia, laying it out
to conserve space better,  or using a smaller typeface. This is
basically identical to the "writing smaller is compression" statement.

* "If you removed the spaces from between the words, that would be
compression." This is known as "The Encyclopedia Whitespace Analogy".

* "If I took a perlscript and removed all whitespace and tokenized the
variables, it would take up much less space on a disk, and indeed be
smaller. This is compression." This is the "Perlscript Analogy".

* "FAT16 did not use drive space as well as FAT32. Since FAT32 can
utilize the space better, it is in a sense, compressed." This is the
FAT analogy.

I think the issue boils down to a few questions. Namely, What is
compression? What is efficiency? What is optimization?

You might say that a perlscript without whitespace and with tokenized
variables is optimized, but is it compressed?

Something that occurred to me about the LD analogy, is that the data
on the outer ring of the LD is *not* compressed -- because the optimal
density for the information is the *centermost ring* of the LD -- in
fact you could even say that by strecthing the signal to fit, you are
enlarging the space the data requires, so that it can remain on one
"ring" of the LD, and that by keeping the data at a fixed density,
you're not compressing it, you're just not enlarging it to fill the
space. So in this case, it is not compression, it's just more
efficient.

Likewise, on the way home last night, I came up with "The Plastics
Molding Analogy". Say you were Mattel, and you were making Barbie
bodies out of plastic, and you used to use a 4"x4" volume of plastic
to mold Barbie's torso. Then, one day, you realized that if I
yin-yanged another torso, I could fit two torsos in ONE 4x4 block of
plastic. I am not compressing the Barbie design, I'm just making my
process more efficient. This analogy is a physical representation of
the FAT analogy -- being more efficient about how you write things
down (without reformatting the design) is not compression, it's just
efficiency, or optimization.

I'll start there, and anyone is free to jump in.

pedro

-- 
----------------------------------------------------------
robotfindskitten.org -- free zen simulation 
xy003.net -- free rock opera
peterpeterson.com -- free skates sharpened and repaired
----------------------------------------------------------

_______________________________________________
UFO Chicago -- Users of Free Operating Systems
Free Software Rules -- Proprietary Drools!
http://ufo.chicago.il.us/cgi-bin/mailman/listinfo/ufo