OT: cd's again... one last question....

[Brian K. White]

----- Original Message ----- 
From: "Mike Schwartz (PC Support)" <mschw at athenet.net>
To: "'Brian K. White'" <brian at aljex.com>; <filepro-list at lists.celestial.com>
Sent: Monday, October 30, 2006 3:52 PM
Subject: RE: OT: cd's again... one last question....


>> I don't. Not even a little. I bet there is no end of things that can be
>> used
>> to match up exact tracks from one edge to another, avan across large gaps
>> from missing pieces, especially with the help of the other not-missing
>> pieces to draw context hints from.
>> I bet it's even easy to have it detect mixed shards from different disks
>> and
>> regroup them appropriately.
>
>> Brian
>
>     The original poster had 100's of data disks to destroy.  My first
> reaction was to tell him that the bits and pieces from each should be
> segregated into several separate piles, then each pile scattered at
> different landfills personally by different trusted people. 
> Reconstructing
> the pieces of 100 disks is way more than 100 times harder than
> reconstructing the pieces of one disk.
>
>     I could offer to scatter them out my airplane window over the middle 
> of
> Lake Michigan.  If the pieces would float, I could scatter them over the
> deeply forested portions of Wisconsin and Upper Michigan that are, for all
> practical purposes, totally inaccessible.
>
>     Reconstructing the pebbles from 100's of destroyed disks would be such
> a horrendous task that nobody would attempt it, unless the data was
> invaluable to recover.  Then again, look at the painstaking effort that 
> has
> gone into the dead sea scrolls and other documents.
>
> Mike Schwartz

What, no Tibetan monks? no full moon? :)


Whats so hard about feeding chips into a hopper that lays the chips onto a 
conveyor belt that passes the chips under a resurfacer, then sets chips one 
at a time on an articulated holder on a disk, spins the disk, reorients the 
chip until the head gets a read, reads the tracks from the chip, databases 
the track fragments and the shape data, repeat next chip....

It takes what a few hours? Days even? to scan all the chips?
Then software just churns away making matches.

I can't imagine the software requires any exotic breakthroughs to write, and 
it's completely easy to just let the software churn over the collected data. 
And hundreds of disks? That's still small numbers even if each disk is 
broken into thousands of pieces even for a cheap desktop computer. Plus the 
numbers get even smaller as soon as it makes it's first match. I have no 
idea how long a single cheap pc would have to work on something like that 
but more cpu and more ram to make the work go faster is the easiest thing in 
the world to provide. I'm sure this kind of work is very paralellable. just 
buy 10 or 20 or 100 diskless pc's and slap together a beowulf cluster.

Physically scattering the chips? Yeah that makes it difficult to reassemble 
of course and even I don't think you can get much use out of a single chip 
or a small enough fraction of the total.
But then the eraser machine becomes more practical.

Plus that point about accountability is a good one.
The other day I was on site at a clients, and the owner printed me an excel 
sheet of all the users login names and passwords for me to enter into the 
backup unix box.
When I no longer needed that sheet, I didn't throw it away or destroy it. I 
brought it back to the owner so that he knew himself what ultimately became 
of it.

Brian K. White  --  brian at aljex.com  --  http://www.aljex.com/bkw/
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro  BBx    Linux  SCO  FreeBSD    #callahans  Satriani  Filk!