Re: too much information!

mensanator_at_aol.compost
Date: 01/14/05 

Date: 14 Jan 2005 13:28:37 -0800

jmfbahciv@aol.com wrote:
> In article <1105647068.240038.96810@c13g2000cwb.googlegroups.com>,
> "mensanator@aol.compost" <mensanator@aol.com> wrote:
> >
> >jmfbahciv@aol.com wrote:
> >> In article <20050112093819.25685.00000061@mb-m26.aol.com>,
> >> mensanator@aol.compost (Mensanator) wrote:
> >> >>Subject: Re: too much information!
> >> >>From: jmfbahciv@aol.com
> >> >>Date: 1/12/05 5:46 AM Central Standard Time
> >> >>Message-id: <PqCdnWhR7560hXjcRVn-hQ@rcn.net>
> >> >>
> >> >>In article
<1105468858.967035.96660@z14g2000cwz.googlegroups.com>,
> >> >> "mensanator@aol.compost" <mensanator@aol.com> wrote:
> >> >>>
> >> >>>jmfbahciv@aol.com wrote:
> >> >>>> In article
> ><1105420562.691975.269220@c13g2000cwb.googlegroups.com>,
> >> >>>> "mensanator@aol.compost" <mensanator@aol.com> wrote:
> >> >>>> >
> >> >>>> >Androcles wrote:
> >> >>>> <snip>
> >> >>>>
> >> >>>> >> 7) Pixels on a computer screen. 16,000,000 colors for each
> >pixel.
> >> >>>> >> How many different pictures are possible?
> >> >>>> >> Is it more or less than a google?
> >> >>>> >
> >> >>>> >For a 640x480 screen, it's 10^2219433, slightly larger than
a
> >> >>>googol.
> >> >>>> But none of it is information.
> >> >>>
> >> >>>I wouldn't say none. A lot of it would be useless. But just
> >because
> >> >>>you can't tell the difference between a picture of a polar bear
in
> >> >>>a blizzard and the inside of a ping pong ball doesn't mean the
> >> >>>picture of the inside of a ping pong ball isn't an accurate
> >> >>>representation.
> >> >>
> >> >>It isn't an accurate representation.
> >> >
> >> >Sure it is. A 640x480x24 array of binary 1s is an accurate
> >> >representation of a uniform white object.
> >> >
> >> >>The picture is only in your imagination.
> >> >
> >> >No, there are two files on my hard drive called
> >> >
> >> >polar_bear_in_blizzard.jpg
> >> >ping_pong_ball_interior.jpg
> >> >
> >> >If the pictures are copied to another disk and renamed
> >> >
> >> >image1.jpg
> >> >image2.jpg
> >> >
> >> >then the context of those images has been lost
> >>
> >> No, it hasn't lost context because there was no context
> >> to lose. "A rose is a rose by any other name."
> >
> >Only if the second file is a copy of the first. If there were
> >two distinct uniform white objects, each object's picture will
> >be a file of all binary 1s and each file is a distinct piece
> >of information even though the content is identical.
>
> You are working with digital devices here. There is no difference
> between the contents of the two files. Thus a program, person,
> or device cannot distinguish between the two. If the task was
> to find the file, all searches would stop at the first one in
> the directory list.

So what? There may be other tasks that don't need to
distinguish the files based on content.

> >
> >Would you want your bank to delete your account because the
> >balance matches that of another customer?
>
> This is precisely why each account has a unique number _as
> an integral part of its data_.

And when it's not possible to put that distinction inside
the file, you put it outside the file. For example:

polar_bear_in_blizzard.jpg
ping_pong_ball_interior.jpg

I've run rings around you logically.

>
> Why do you think the post office looks at the last line of
> an address label first? You clearly need to learn a lot more
> about sorting and binning.

Been there, done that. My first job was with
Cummins-Allison. We made check reader/sorters for
the banking industry.

> Get thee a box of IBM cards, punch data, find a sorter,
> collator and merger,

Luxury. I had to work on those goddamn paper tags
they used to use in clothing stores. Punch card readers
were a doddle. You haven't lived 'til you've debugged
a tag reader.

> then play with all
> machines, noting all the different arrangement of cards you
> can have.

Suppose (as is more likely) all you have is the
punch cards. You could recover the information off the
cards by manual inspection or, by reverse engineering,
build a reader.

Now reverse it, i.e. you have a sorter, collator and
merger but no cards. Now what do you do? Data without
context is better than context without data.

> >
> >>
> >> > .. and they
> >> >become useless. But they remain valid pictures.
> >>
> >> They are not valid picture_s_. It is one, and only one, picture.
> >
> >Only if they were created invalidly. If I opened Windows Paint
> >and saved the default white canvas and then _claimed_ it was
> >a photo of the interior of a ping pong ball, then you could
> >claim it was an invalid picture.
>
> It is indistinguishable from white noise;

Actually, white noise wouldn't look anything like that.
It would look more like the snow you see on your TV when
it's tuned to an unused channel.

> thus it's garbage data. Now there are uses for such things,

Aha!

> but recognition based on its contents is not one of them.

But you originally said "none of it is information".
So if there are uses (aside from content recognition),
then there is, in fact, _some_ information.

"And when I say there's none,
I mean there is a certain amount."
- Monty Python

> >
> >It may be that no two objects are _exactly_ identical, but they
> >don't have to be to take the same picture. The camera has a finite
> >resolution and light sensitivity, so two distinct objects may
> >appear identical to the camera.
>
> Take two pictures of the same thing. Do a binary FILCOM of the
> two files. You can test your ideas. Another word you can
> probably google to find out more is granularity (at least,
> that was one of the cybercruds used waybackwhen).
>
> >
> >>
> >> > .. The information
> >> >has not been lost, only the context in which it was used has.
> >>
> >> There is only one piece of information with this picture and that
> >> is the color to paint the TTY screen.
> >
> >Now you're starting to sound like Androcles. It's one piece of
> >information PER PIXEL. And there are 307200 pixels.
>
> No.

"No" you're not talking like Androcles or "No" there aren't
307200 pixels?

> Now you are talking about how to display the bits that
> are stored. Your pixel map will not necessarily produce the
> same display on my TTY as it does on yours.

But it doesn't matter whether it looks the same.
You said there one piece of information - color.
That implies that all screen sizes have the same
look (even if the look is wrong). But that's
incorrect. If the TTY screen has fewer pixels than
the picture, the picture will be cropped. The pixel
count is information in addition to the color.

>
> > .. It is
> >irrelevant to this discussion that every one of those 307200
> >pieces of information is identical.
>
> I give up.

That's to be expected when you're arguing from a
position of weakness.

> I hate cute.

And yet, you say such cute things as
"GIGO is still GIGO no matter what you name it."

> And you know not to play the cutseys from our last discussion.

Was that the one where you called me an ***
for making light of your paranoia about usenet
trolls?

Maybe I enjoy being abused by Mistress BAH.
>
> <snip>
>
> /BAH
>
> Subtract a hundred and four for e-mail.