The Weight of an e-Book

whoever57 writes "According to Prof Kubiatowicz from Berkeley, each time an additional book is downloaded to an e-reader, the mass of the e-reader increases. The effect doesn't really make the devices more difficult to carry: the professor calculates that 4GB of books would increase its weight by a billionth of a billionth of a gram— about the mass of a single virus or DNA molecule."

Comments

So it turns out....

[TheInternetGuy]

So it turns out, pirating is stealing after all?

Re:So it turns out....

[siddesu]

Not if you pay for the electricity that is used for all these writes onto the media.

Re:So it turns out....

[TheInternetGuy]

So if I understand this correctly. If I were to copy , (using the cp or copy command) some copyrighted files from friends PC to my USB.

That would be mere cloning / copyright infringement?

While if I were to use the mv or move command that would be stealing?

OK, got it. Thanks.

Re:So it turns out....

[TheInternetGuy]

My mom pays for all basement electricity, you insensitive clod!

Re:So it turns out....

[TheInternetGuy]

You mean, she knows I'm down here?

Re:So it turns out....

[TheInternetGuy]

Is that some new brand of wooden peg legs?

Re:So it turns out....

[beelsebob]

Not only that, even if it were deleted, then the electrons that make up his copy are clearly different electrons. He's not stolen, he's copied, and then caused criminal damage.

Re:So it turns out....

[Anonymous Coward]

#!/usr/bin/bash

# steal.sh

cp $1 $2

dd if=/dev/zero of=$1

Re:So it turns out....

[rgbatduke]

Ah, but the information theorist would argue -- with absolutely impeccable reasoning, mind you -- that both empirically and theoretically, information is strictly conserved in all physical interactions: no new information is ever created. In fact, existing physical laws strictly conserve information. Leonard Susskind has a lovely book about his "Battle with Stephen Hawking to make the world safe for quantum mechanics" that demonstrates that even black holes conserve information (and even if they didn't, nobody ever argued that they create it.

Consequently the information content of any book was, in fact, not created "by" anyone. It was inherent in the Universe itself from the non-existent oxymoronic "beginning of time" (or rather, the four-dimensional Universe itself is a static entity unless/until proven otherwise). Therefore, one cannot "steal" it (or "pirate" it) as it was not created by the author -- that is strictly an illusion -- and indeed the so-called "theft" is as inevitable as the original appearance of the bizarre and highly unstable patterns of electronic, molecular, and neural state that represent the ontological entities we call "a book".

Damn shame about the inevitable and eternal appearance of those same evanescent patterns that represent the "DMCA", wot?

rgb

Re:The most pointless /. post evar...

[TheInternetGuy]

Not really.
It prevented me from posting "First Post!!! suckers!!!" and instead post something semi-meaningful.

Is this news?

[neonv]

I want my 30 seconds back ...

Not true at all.

[Anonymous Coward]

From TFA:

Although the electrons were already present, keeping them still rather than allowing them to float around takes up extra energy – about a billionth of a microjoule per bit of data.

No matter whether any bit is currently being used or not, it still has a value. It's not allowed to "float around".

Re:Not true at all.

[niftydude]

I think they're talking about flash memory, which does involve confining excess electrons in an isolated piece of material (the "floating gate") to produce zeros.

Yes - but if the flash memory is formatted - it would be all zeros already.

So if the book is downloaded - all the extra ones created should release the excess electrons, and actually make the book lighter!!!

However - if the memory was quick formatted - unwritten memory would be randomly 1 or 0, and adding an ebook would typically keep the same average of 1s and 0s - meaning the reader would on average stay the same weight.

Re:Not true at all.

[rdebath]

Flash memory cannot be 'quick formatted' each block has to be properly erased before use because writing can only turn 1's into 0's. (Obviously, the filesystem on a flash device can be quick formatted; but a recent OS will tell the flash about this, using the "TRIM" command, and the flash will erase all the blocks anyway.)

Some flash drives even understand the NTFS filesystem well enough to erase unallocated blocks without help; but that seems a little dangerous to me.

BTW: What Kubiatowicz seems to be saying is that pulling electrons from the substrate into the gates of a flash drive makes it heavier. So erasing the blocks, ie shorting them to ground, makes them lighter. So while downloading a book could make the device lighter, erasing the device will make it lighter still.

Re:Not true at all.

[rgbatduke]

Actually, I was about to post the same, but then I thought about it. Suppose the medium is magnetic. "Erased" it is in a homogeneous state, which is actually a lower energy state than when there are patterns of 0's and 1's on it. Every 0/1 boundary costs energy. For magnetic media, then, writing any detailed data to the drive costs energy (most of which is expended shifting the state in a bit in a way that conserves the final energy of the state in a symmetric way) BUT there is a very slight increase in the energy associated with neighboring bits when they conflict. As you use the drive you approach an equilibrium average energy slightly higher than all 0's on the drive and slightly lower than a drive full of 01010101...s

For flash memory (if I understand it correctly, open to doubt) a bit is basically a capacitor with an RC time constant of years and a double-gate MOSFET hooked up that permit the passive reading of capacitor state. The default, energetically ground state of a bit is 1. Writing a 0 to a bit "flashes" electrons across a quantum boundary to invert the capacitor state, and the inverted state is very slightly higher energy (and increases the interbit energy as well in a manner similar to the HDD platter).

Note that one should not confuse (in either case) the energy required to switch a bit's state with the energy of the final state. The former is quite large but doesn't add energy/mass to the storage medium. I'd have to read TFA to see if the author actually confused this, but if I read the article then my head would probably explode and I have other more useful things to do -- like almost anything. Indeed, the design of flash memory looks like it might well be (almost?) completely energy-symmetric -- I'm really assuming that the 1 state is energetically stable because of the way the device is hooked up to ground -- the actual structure of a bit is two N-type "plates" separated by a P-type insulator/substrate and looks completely symmetric. If the true ground state is symmetric neutrality, then the author is incorrect -- flash memory, as you say, has the same energy over ground once initialized/formatted in ANY state, and the only increase in energy is a second order interbit interaction energy. For ferromagnets that energy would be minimum in the aligned state. For flash memory, I'd actually expect it to be antiferroelectric -- lowest energy when an equal number of 0s and 1s are written to the device, so that pairs of bits could neutralize each other and enter not-0 not-1 with no net charge (asymmetry) on either N-type plate. But I rather think that 1 is already this neutral state.

rgb

Re:Not true at all.

[rgbatduke]

Just to be picky -- the default state is 1's, not 0's. NAND in particular has to start out all 1's and then "writes" turn some bits in a block into 0's. And the issue is whether or not the 1/default state is the ground state of any given bit (first) and whether or not there is some interbit interaction energy (second) and what the sign of that interaction energy is (third). One could in principle write a very simple model hamiltonian for the system that looks like:

H = - \sum_i (A b_i +/- \sum_j B_{ij} (b_i - 1/2) (b_j - 1/2))

where the first term represents the additional energy gained turning a 1 into a 0 and the second one (probably summed only over nearest neighbors) the energy gained or lost when neighboring bits are the same or different states. b_i is bit state, 1 (default) or 0.

The real question, then, is whether or not A is zero or if it should be e.g. A(b_i - 1/2) (symmetric) and whether B_{ij} is positive, negative or zero. If I didn't make any algebra mistakes in writing this down. If A and B are known, of course, one can easily estimate the cost of writing 4 GB of data, and I'm guessing that's what TFA does (without reading it, of course, what would be the fun of that!).

rgb

this is the most retarded thing i've ever read

[mug funky]

in other news, ipods get heavier as you fill them.

maybe "the singularity" will happen when the internet gets so heavy the Earth collapses into a black hole?

Re:this is the most retarded thing i've ever read

[Samantha Wright]

Your ideas are intriguing to me and I wish to subscribe to your newsletter.

Real units?

[GreennMann]

"billionth of a billionth of a gram" That is painful to read. How about scientific notation? 1*10^-18 grams Or the use of a prefix? 1 atto gram

Re:Real units?

[whoever57]

Blame samzenpus. My submission said 1e-18.

Re:Real units?

[perpenso]

"billionth of a billionth of a gram" That is painful to read. How about scientific notation? 1*10^-18 grams Or the use of a prefix? 1 atto gram

I fear you have just validated the original author's choice. :-)

Re:Real units?

[ari_j]

In a legitimate publication, the editor is a person responsible for ensuring factual accuracy and good writing style before an article is published. On Slashdot, an editor is a person responsible for taking factually accurate, well-written submissions and ensuring that they lose at least one of those attributes before publication. So we can't really blame samzenpus--he's just doing his job.

Where's the actual claim?

[White Flame]

This just links to a Telegraph article talking about something that was talked about in a New York Times article, with no link to either that or the original source. Come on, Slashdot.

Science?

[No, I am Spratacus!]

This belongs in the Idle section, at best, but probably not at all on /.

Re:Wait a minute...

[ttong]

That's not how woosh works, you don't woosh your own reply.

Related article:

[6Yankee]

Under TFA: "Amazon Kindle review: the e-reader for the mass market"

Re:Hmm...

[maxwell demon]

There is a theoretical limit on the amount of energy you need to erase a bit (in order to store new information in). It depends on the temperature, though. At room temperature it's about 2.8×10^(-21) joules, or the energy equivalent of 3.1×10^(-38) kilograms.

Re:0 is heavier than 1?

[Tuqui]

With the same reasoning one could argue that batteries increase in mass when charged as they take up energy - a probably much larger quantity of mass as the energies involved are so much greater.

But when you charge batteries the temperature of the appliance will climb few points. If we take account of the molecules of air displaced out of the appliance by the raising in the temperature an expansion of the air inside, the total weight of the appliance will decrease.

If you read the source article at NYT...

[xded]

... (which the editors should've linked to), it states:

“Although the total number of electrons in the memory does not change as the stored data changes,” Dr. Kubiatowicz said, the trapped ones have a higher energy than the untrapped ones. A conservative estimate of the difference would be 10^(-15) joules per bit.

As the equation E=mc^2 makes clear, this energy is equivalent to mass and will have weight. Assuming that all these bits in an empty four-gigabyte Kindle are in a lower energy state and that half have a higher energy in a full Kindle, this translates to an energy difference of 1.7 times 10^(-5) joules, Dr. Kubiatowicz calculated. Plugging this into Einstein’s equation yields his rough estimate of 10^(-18) grams.

Of course Kubiatowicz also says that:

[10^(-18) grams] is only about one hundred-millionth as much as the estimated fluctuation from charging and discharging the device’s battery.

Which is a far better comparison than the one obtained from The Guardian where Graeme Ackland of Edinburgh University stated:

"If Prof Kubiatowicz is really struggling with the extra weight, he is welcome to come to Edinburgh where it's cooler, and the lack of thermal energy in his Kindle will more than compensate."

Slashdot, home of crowdediting.

Re:Wait a minute...

[GrumblyStuff]

You mean the iTruck?

Don't tell the cheap airlines

[asdf7890]

I can see RyanAir and friends using this as an excuse to add a new "eBook reader carrying charge" to all flights.

Re:The most pointless /. post evar...

[Chrisq]

Really...

I agree. It was not worth the extra weight that downloading it added to my laptop

Re:Harddrives?

[TheRaven64]

He seems to be under the impression that the storage devices have three states: one, zero, and undefined. This is not the case. There is no undefined state, when flash is erased all of the bits are set to one, when it is written some are set to zero. There is no difference in energy state between the a block that is erased and a block that is storing all ones. It is possible that the zero energy state has more mass than the one energy state, but that's not what he is claiming. Expect to see this show up in The Guardian's Bad Science column soon...