How Do I Nuke My Hard Drive?

[Beady]

I've finished moving into my new computer. The old one is going to charity. How do I nuke the hard drive? I can't do it from inside Windows, including from the command prompt in the DOS window.

I know there's a way to boot into DOS, but how do I manage it?

[Darat]

Beady: Set of instructions can be found here (http://www.computerhope.com/issues/ch000186.htm)

[Beady]

No joy. This is a one-time operatiojn, so I don't want to have to pay, but the freebies aren't for XP or don't want to download.

[CFLarsen]

"Nuke", as in completely destroy all data there?

A. Big. Fat. Magnet.

[LordoftheLeftHand]

I think what you’re looking for is a program that will perform a binary wipe on your whole hard drive, which comes with a boot image (the os of the image wouldn't really matter). If you can't find one by this evening I'll see what I can dig up.

LLH

[Smike]

Make a windows ME bootdisk (http://www.bootdisk.com/bootdisk.htm).

Boot it up, select any of the starting options.

When you get the DOS prompt, type "format c:" (where c is your hard drive) and hit enter

If you have partitions you want to remove, type "fdisk"

Unless you're looking for a extreme security wipe...

[jnelso99]

http://dban.sourceforge.net/

[Anti_Hypeman]

If you really want to give it away just buy a super cheap 20 gig to stick in there. You can get used HD's for chump change on eBay its worth it for the peace of mind, you never know whos going to end up with it. I suspect 2nd hand computer shops are into identity theft.

[FFed]

The hard drive manufacturer should have some tools to download which can do that.

[HeyLeroy]

go to pcworld.com and look for BCWipe.

[Beady]

go to pcworld.com and look for BCWipe.

That's considerably expensive, and I want this for only a single operation. I did, however, find a freeby on that site called "Killdisk," which seems to work. It's running as we speak, and appears to be doing the job.

Thanks, to all who answered.

[Molinaro]

If you have any version of a Windows install CD, boot off that disk and start the setup process.

When you get to the point of selecting what partition to insall on, you will be able to delete the current partion, create a new one, and if you want.. format it.

Then just abort the install.

[Molinaro]

Alternatively...

During bootup, press the F8 key to get to the startup option menu. There (depending on what version of Windows) you will see a "Restart in MSDOS mode".

From there you can either Fdisk to kill the partition or just Format it.

[bignickel]

I'm not sure some of the suggestions listed above will work (the non-buying-software ones). Remember: unless you ACTUALLY overwrite each space on the hard drive with a 1 or 0 (either randomly generated, or ancient Latin text to raise the dead), then whatever was on the hard drive, will STILL be there, even if no tables points to it. And a hard drive reading program could still recover it.

The only way to destroy data on a magnetic medium is to overwrite with something else.

[jnelso99]

From what I've heard, just overwriting the data once will not do the trick, especially against folks like the NSA. I had used Autoclave before (http://staff.washington.edu/jdlarios/autoclave-discontinued/), and in the FAQ there's a blurb about the various erase levels:


What's with the levels 1 through 5? Which level do I want to use?

All levels erase the disk completely. The only difference is how difficult it would be for someone to recover data from the disk using sophisticated recovery tools (including scanning tunneling electron microscopes). Level 1 is the fastest, level 5 is the slowest. Level 5 is the most secure, level 1 is the least secure. I personally couldn't recover anything from a disk that had been cleaned with level 1, but someone with the know-how and a few thousand dollars could. I'm not guaranteeing anything, but I doubt the NSA could recover anything from a disk that had been cleaned with level 5. Level 3 meets most corporate and nonclassified government erasure specifications. Here's what each level does:

* 1 - A single pass of all zero.
* 2 - One pass of random data followed by one pass of all zero.
* 3 - Three passes: all zero, all one, all zero.
* 4 - Ten passes, some of which are random, followed by one of zero.
* 5 - 25 passes, three of which are random.

Note that none of these levels verifies that data is actually getting written to the drive. That will be a future addition, as will a level which conforms to DoD specifications.

[Molinaro]

I'm not sure some of the suggestions listed above will work (the non-buying-software ones). Remember: unless you ACTUALLY overwrite each space on the hard drive with a 1 or 0 (either randomly generated, or ancient Latin text to raise the dead), then whatever was on the hard drive, will STILL be there, even if no tables points to it. And a hard drive reading program could still recover it.

The only way to destroy data on a magnetic medium is to overwrite with something else.


Since the OP says nothing about wanting to overwrite every byte on the drive, why jump to that conclusion?

Do you honestly expect the charity on the recieving end to go all James Bond with trying to read it?

Further evidence that the OP was not refering to an attempt to overwrite every byte would be that he mentions having tried to do it from within a DOS prompt, and that he was trying to boot into DOS. That is highly suggestive that what he was trying to do was execute a Format command on the active partition.

[LordoftheLeftHand]

from that article:

* 1 - A single pass of all zero.
* 2 - One pass of random data followed by one pass of all zero.
* 3 - Three passes: all zero, all one, all zero.
* 4 - Ten passes, some of which are random, followed by one of zero.
* 5 - 25 passes, three of which are random.



I've seen free utilities that will do 3 passes (0,1,0). Couldn't you just run that several times to achieve higher security?

LLH

[kevin]

Since the OP says nothing about wanting to overwrite every byte on the drive, why jump to that conclusion?

Because the OP may not understand that killing the partition and doing a simple format c: frequently doesn't actually wipe any data and its easily recoverable. It's a good idea for this type of question to recommend a total wipe AND explain why.

Do you honestly expect the charity on the recieving end to go all James Bond with trying to read it?

Nope, don't expect it. But I've seen it. Many charity organizations have voluteer help that like to sniff around just for the heck of it when they get bored. Many donations go through many hands before finally landing somewhere, i wouldn't trust everyone in that delivery chain.

Better safe than sorry, and it's not like this is an expensive option there are many free tools available.

[kevin]

from that article:



I've seen free utilities that will do 3 passes (0,1,0). Couldn't you just run that several times to achieve higher security?

LLH

Several passes of random writes is best, but more than 3 probably unneeded. If you're really that paranoid about the data on the drive that 3 passes won't protect you're best bet is to completely destroy the drive (open that sucker up and smash the platters).

[Molinaro]

Write 0 to all, then 1, then 2, then 3, up to 11111111..11 in binary for whatever size you're writing, then 0s again.

[kevin]

Write 0 to all, then 1, then 2, then 3, up to 11111111..11 in binary for whatever size you're writing, then 0s again.

If the sequence of numbers written is known then it is possible to figure out the original data that was overwritten (within certain percentages of likelyhood).

For example, if a bit was a 1, and is rewritten to a zero the magnetic field doesn't go all the way to what is identified as a zero. if you read the strength back, and know that it was supposed to be modified to a zero you can figure out which bits were 1's prior to the change and which were 0's. The more passes made the more likely the original field gets wiped. If the bit pattern of the overwritten data is known, it makes it easier to figure out what the original data was and more passes are needed to obliterate the original data.

Your example useses a lot of passes, so it would work, but it would take a long time.

3 passes of reasonably random data would be much faster and most likely just as effective.

[LordoftheLeftHand]

Thanks for the info Kevin and I agree: If it is that important a sledgehammer might be a better solution.

Still very interesting.

LLH

[jimlintott]

It isn't mil spec but here (http://www.objmedia.demon.co.uk/freeSoftware/secureDelete.html) is a free one for windows.

[Anti_Hypeman]

In the National Guard we have to remove the HD and send it to be destroyed before getting rid of the computer. When I worked on a NATO base we had a incinerator. The military dosent use these military grade shredders so I dont know how good they really are. If you ever had important information on it its not worth the risk when you can replace it for peanuts. Why not just donate it without a HD?

There is a chance the charity will just sell it to a 2nd hand shop. It very well could end up in the hands of a super evil super geek.

[RichardR]

I shred (http://www.pcmag.com/article2/0,1759,13352,00.asp) my files.

Free download (http://www.pcmag.com/article2/0,1895,943259,00.asp)

[Zep]

One to add to the list: Seagate Zerofill - http://www.seagate.com/support/kb/disc/faq/ata_llfmt_what.html

[bignickel]

When I was googling on 'hard drive permanent file removal' or such, I came across the web site of someone selling their nuking tool.

He had an anecdote on his front page about, surprise: guy donates computer to charity, and then huge purchases start showing up on his debit card. Had someone obtained his hard drive from the charity? Yes, they had.

When it doubt... don't give them the hard drive. They can use one of their own, or a new one donated by a hardware company.

[Soapy Sam]

It's worth taking a hard drive apart, if only for the magnets. (I admit to being a magnet freak).
And the disk itself makes a super pocket mirror.


OK. OK. Big nerdy kid.

If you did have any data on there you don't want loose, like the firing codes for a polaris missile or something, wrecking the drive is the only absolute 100% way to be sure. If you don't have a screwdriver, a hammer and chisel will do.

At the OS level, why would this not be enough?- Format the drive. Reinstall the OS. Write a script that simply copies itself (S1.vbs, S2.vbs etc) to the drive. Make it the right size to nearly fill one sector, so there's no unwritten space. Then let it run till the disc is full?
I wonder how long a script would take to fill a 60Gb drive?

[Beady]

From what I've heard, just overwriting the data once will not do the trick, especially against folks like the NSA.

This isn't that big a deal. I gave the thing to a charitable recycling shop that we've donated to several times. If the person who winds up with my computer could do half the things you guys are talking about, he wouldn't be shopping there.

[Zep]

At the OS level, why would this not be enough?- Format the drive. Reinstall the OS. Write a script that simply copies itself (S1.vbs, S2.vbs etc) to the drive. Make it the right size to nearly fill one sector, so there's no unwritten space. Then let it run till the disc is full?
I wonder how long a script would take to fill a 60Gb drive?Nope, not secure at all. There's a ton of space on most HDs that is reachable that is not necessarily erasable by any standard OS tools or methods. Zapping a disk needs to hit stuff like PC BIOS storage areas (sometimes up to 2MB), multiple boot blocks, track 0 space, bad block space, etc, etc (i.e. where the virii live...).

As said above, you need a really low-level formatter - one that literally re-applies the blocking information, rather than just rewriting the existing block contents. Or drive a stake through its heart!

[Soapy Sam]

Nope, not secure at all. There's a ton of space on most HDs that is reachable that is not necessarily erasable by any standard OS tools or methods.-Zep

Is it likely that any of Beady's dirty pictures are stored in those areas though?
Surely anything stored by the OS can be overwritten by the OS?

Beady. Keep the damn thing as an internet terminal. That way you don't have to worry about malware, just reformat it twice a week.

[Pyrrho]

Where I work, we use "Eraser" to completely erase files and free space on hard drives.

http://www.heidi.ie/eraser/

Freeware. Works on XP.

[kevin]

At the OS level, why would this not be enough?- Format the drive. Reinstall the OS. Write a script that simply copies itself (S1.vbs, S2.vbs etc) to the drive. Make it the right size to nearly fill one sector, so there's no unwritten space. Then let it run till the disc is full?
I wonder how long a script would take to fill a 60Gb drive?

That'll work fine for data you're most worried about (i.e. your personal data). You could even write a vbs script to generate random numbers and write them to a file on the disk, then you don't have the issue of known bits being written to the drive. Heck it's probably slightly faster since the drive heads don't have to move to read data before copying it.

However the secret is always multiple passes. Make sure you do this at least 3 times.

Oh yeah, and if you plan on tossing a drive in the trash you should probably do this too. But then, I shread my junk mail as well as my regular mail.

[bigred]

I give up: do you have some Windows licensing issue here? ie why go through all this "kill HD" thing and make it basically un-unusable to a charitible organzaition, who would now have to go BUY Windows to use your "free" gift?

Why not just delete everything and then do a defrag for ex?

[Smike]

I give up: do you have some Windows licensing issue here? ie why go through all this "kill HD" thing and make it basically un-unusable to a charitible organzaition, who would now have to go BUY Windows to use your "free" gift?

Why not just delete everything and then do a defrag for ex?

If you mean not unistalling windows, that would be illegal if Beady wants to use the same copy elsewhere.

There are also security issues, but unless Beady has been keeping very compromising data on his PC, I wouldn't think that that would be a problem.

[Beady]

If you mean not unistalling windows, that would be illegal if Beady wants to use the same copy elsewhere.

There are also security issues, but unless Beady has been keeping very compromising data on his PC, I wouldn't think that that would be a problem.

Look, guys, let's get a grip. The place I gave it to, they do their own nuke and reinstall, anyway. In fact, the computer will probably get a good servicing and upgrading, and be indistinguishable from a braned-new, if very low-end, machine. It will probably also sell for less than $100, because that's about the income-level of people who shop there.

As for Windows, it'll have a brand-new, perfectly legal, install. I don't know how, but these recycling shops are tied into a supply network that no one like us normally sees. I'd be willing to bet that MS donates licenses to these places. We got to know this particular place pretty well during our decade of unemployment.

FWIW, I wouldn't know how to buy a computer that didn't have Windows already installed. My big problem with the new one is figuring out what installed stuff I don't need and how to get rid of it.

[Smike]

In that case, either delete the stuff you don't want the charity people to see, or if there's loads of it, format the hard drive, using a bootdisk as I suggested earlier.

[Beady]

In that case, either delete the stuff you don't want the charity people to see, or if there's loads of it, format the hard drive, using a bootdisk as I suggested earlier.

Already done. No one seems to have noticed, but I've been writing in the past tense for the last couple of days.

[LordoftheLeftHand]

who would now have to go BUY Windows

Who actually buys windows products?

LLH

[Beady]

Who actually buys windows products?

For the computer I just got rid of, I bought something called Office Suite, so I could get MS Word, rather than have to use the "Microsoft Works Word Processor" that came pre-installed. Weird thing was, I could have paid $400 for Word by itself, but the Suite bundle was $150. The even weirder thing was that Word was the only part of the bundle I ever used. Upshot was that I paid $150 for a bundle of MS products that I didn't want, so I could have a program I would have had to pay $400 for, all on its own.

I understand MS makes money on deals like this.

[Smike]

I understand MS makes money on deals like this.

They sold you a product that is available free (i.e. OpenOffice). They charged you $150.

Yes, they probably make quite a lot of money.

[Zep]

Ssssh! I know some people in here who work for Bill!

[Gromit]

Ironically for this forum, all the talk about more than one overwrite pass needed to totally wipe data is actually anecdotal and has never been shown to be true. I'd post a link to a page I wrote discussing this topic but...first post, can't post links - you know the drill.
The idea that you need multiple numbers of overwrite passes is based on an old paper written on theoretical data recovery on hardware made about 10 years ago or more. As you can imagine, very few hard drives are still in use today that were made a decade ago.
If it helps, as a long-time lurker I finally registered just so I could speak on this topic.
For nuking disks, I'd recommend a DBAN bootdisk. A single overwrite pass with that will mean that no-one on the planet is getting your data back.

[Soapy Sam]

spell it out. Someone else will post it.
Or make 14 pointless posts.

Oh- and congrats on decloaking.

[kevin]

Ironically for this forum, all the talk about more than one overwrite pass needed to totally wipe data is actually anecdotal and has never been shown to be true. I'd post a link to a page I wrote discussing this topic but...first post, can't post links - you know the drill.

post without the http:// on it. we can copy/paste the url.

Here's a link to the original paper I believe you're referring to:
http://www.cs.auckland.ac.nz/%7Epgut001/pubs/secure_del.html

It mainly deals with recovering data from MFM or SPM encoded drives (no longer used). In the epilogue the author talks about more modern drives:

"For any modern PRML/EPRML drive, a few passes of random scrubbing is the best you can do. As the paper says, "A good scrubbing with random data will do about as well as can be expected". This was true in 1996, and is still true now."

Not sure when the epilogue was written.

For nuking disks, I'd recommend a DBAN bootdisk.

DBAN available here:
http://dban.sourceforge.net/

[Gromit]

Here's a link to the original paper I believe you're referring to:
[Gutmann paper link]

Well, I _do_ refer to that rather elderly article. But the page in question is shsc.info/DataRecovery
What I thought was ironic about the discussion here is that there is no physical evidence of anyone on the planet ever manging to recover a useful amount of overwritten data. And by "useful", I mean anything more than a few bytes. Reliability of recovered data is woeful. If anyone knows of anyone other than Professor Gomez at the Uni of Maryland, please let me know. He seems to be the only one who has bothered cracking out the STEM to give it a go.
So to me, as someone who has a vested interest in the field and has done rather a lot of asking around on the topic, it is very similar to dowsing ;)

[kevin]

clickable version of your page:
http://shsc.info/DataRecovery

no disagreement that the software tools listed on your page won't recover a single pass, but I think (with, as you point out, no real proof yet) hardware forensics tools could recover from a single overwrite, especially when overwritten with known data.

In digging around a bit I've found an interesting bit of trivia. You ask "So why does the Department of Defense specify that huge multi-pass overwrite if one is enough?"

As far as I've been able to find the DOD doesn't actually recommend multi-pass overwrite. They only allow destruction.

The document cited is frequently: DOD 5220.22-M
Which can be found online here:
http://www.dss.mil/isec/nispom.pdf

See chapter 5, section 7 (page 51 of the pdf). No mention of multiple passes of hard drives that I can find. Just destruction.

[Molinaro]

Being someone who *hates* opening .pdf links.. I'll offer a nice .htm page that includes a nice little summary matrix of the Military Specs for various storage mediums:

http://www.killdisk.com/dod.htm

[kevin]

Being someone who *hates* opening .pdf links.. I'll offer a nice .htm page that includes a nice little summary matrix of the Military Specs for various storage mediums:

http://www.killdisk.com/dod.htm

HTML versions (and Word doc versions if you're a masochist) of the whole thing are here:

http://www.dss.mil/isec/nispom.htm

Now what's odd is that cleaning/sanitizing matrix is in the 1995 version, but seems to be removed from the updated version. I don't think that is part of official policy anymore. And the old matrix itself states 3 passes is sufficient but shouldn't be used for really secret things.

And as Gromit points out there isn't any real evidence (that I've found) of one-pass drives being recovered. I'll stick with my 3 passes as it isn't really that much more. I never bought into the 20 pass crap anyway.

[Gromit]

That's the unsual thing, isn't it. On a sceptic web forum, where proof is king, people still hold onto the "more than one overwrite is needed" thing even though no evidence shows it.

kevin: No, there is no hardware recovery solution available. The best people in the world to handle it is probably Kroll OnTrack - they are who the US military and the NSA use for data recovery. I hate to think what hurdles they needed to jump through in order to get the clearance.
But they can no more recover once-overwritten data than you or I. As a computer forensic analyst, I'm in contact with an international body of people with an interest in this field and no-one has been able to find ANY evidence of this kind of recovery being possible.
Given that, why do people cling to this enduring belief? (Or rather, when asked to precure evidence and fail to do so.) Why is it treated differently to that of more classicly paranormal claims? Just because it relies on fantastical statements based on real physics, rather than blatant pseudo-scientific claptrap?
I'm not expecting to change people's minds here, much the same as a lot of posters here don't think that they will change a woo's opinion, I guess. Excuses will always be forthcoming, the most popular, of course, being "the guvmint can do it, but they don't tell no-one."
Don't get me wrong - I WANT this to be a reality. I have several homicide cases that could do with the information. I'm not holding my breath, however.

[sparks]

If your issue one of security, then just remove the HD and run a drill bit through it at high speed. If that's too much like work, give it a 'Clint Eastwood' wipe.

[jimlintott]

I'm not expecting to change people's minds here, much the same as a lot of posters here don't think that they will change a woo's opinion, I guess. Excuses will always be forthcoming, the most popular, of course, being "the guvmint can do it, but they don't tell no-one."

I never really gave it much thought before as it wasn't much of an issue to me personally but you are right. A little research shows that there is no evidence of extreme data recovery working. One article I read discussed the electron microscope idea and clearly showed that it could take years to recover even a small amount of data.

[Ed]

I'm not expecting to change people's minds here, much the same as a lot of posters here don't think that they will change a woo's opinion, I guess. Excuses will always be forthcoming, the most popular, of course, being "the guvmint can do it, but they don't tell no-one."
Don't get me wrong - I WANT this to be a reality. I have several homicide cases that could do with the information. I'm not holding my breath, however.

Interesting points. If NSA could decipher erased data more effectively than you are suggesting it would seem to me to be worthy of the highest level of security classification. I would also be worthy of a subtle and confusing disinformation campaign, no?

[Ceritus]

I've finished moving into my new computer. The old one is going to charity. How do I nuke the hard drive? I can't do it from inside Windows, including from the command prompt in the DOS window.

I know there's a way to boot into DOS, but how do I manage it?


Do a keyboard error and then go to the bios screen and exit out of it without going to windows and format from that dos prompt.

Or just simply go to your bios screen selection if you know how to...

[Gromit]

Interesting points. If NSA could decipher erased data more effectively than you are suggesting it would seem to me to be worthy of the highest level of security classification. I would also be worthy of a subtle and confusing disinformation campaign, no?
Except, like I said, the NSA outsource their recovery to a private company. A private company that is only too aware of the MASSIVE financial gains to be made from the ability to recover data in this manner.
Still, if you feel the tinfoil hat fits then feel free to wear it ;)

[kevin]

Given that, why do people cling to this enduring belief? (Or rather, when asked to precure evidence and fail to do so.)

Their are different levels of "cuckoo" science. You've equated my statements to that of dowsing. I think a better comparison is to the Vitamin C claims.

Everyone agrees that Vitamin C is necessary. You get any and you get scurvey. But what about claims that taking way more prevents cancer, heart disease, etc....

With Vitamin C (or dowsing) claims I can point to a body of peer-reviewed scientific studies showing it does NOT work.

Now we both agree a single-pass data wipe is necessary. Now what about claims that more passes are need to ensure the data can't be read back.

I can point to one paper whose author claims multiple passes are still needed for drives.

The best you can provide is what lawyers call hearsay evidence that would not be presentable in court, or usable in a peer-reviewed paper.

The ONLY difference between our two methods is that mine takes 3 times longer. Until you can provide real EVIDENCE, not hearsay, of someone trying this, what equipment and methodology they used then I'll have to stick with my way.

And I've presented no supernatural claims. My claim is that when a magnetic field that is changed, their may still be evidence of it's previous state. This is a known fact of physics. The question is, with the current methods used to encode data on the drive is there enough evidence on the drive to recover this previous state. And if there is, how many times do you have to change the state in order to remove that evidence. The question comes down to one of technological ability, not supernatural claims.

[kevin]

If your issue one of security, then just remove the HD and run a drill bit through it at high speed. If that's too much like work, give it a 'Clint Eastwood' wipe.

this only destroys data in the immediate area of the hole. most of the disk is still recoverable. Shattering (in enough pieces to make reassembly difficult) or melting the platters is preferrable for drive destruction.

[jimlintott]

this only destroys data in the immediate area of the hole. most of the disk is still recoverable. Shattering (in enough pieces to make reassembly difficult) or melting the platters is preferrable for drive destruction.

Really, I would have thought that all the debris and damage caused by drilling would take out the heads in a heart beat. Isn't it almost impossible to move platters from one drive to another due to them needing extremely precise alignment?

[Gromit]

Now we both agree a single-pass data wipe is necessary. Now what about claims that more passes are need to ensure the data can't be read back.

I can point to one paper whose author claims multiple passes are still needed for drives.

The best you can provide is what lawyers call hearsay evidence that would not be presentable in court, or usable in a peer-reviewed paper.
Actually what I'm pointing out is your "evidence" is a purely theoretical paper that is based on technology that is not in use any more. You also have ZERO evidence of this method working on current equipment, whereas there are a variety of documents written by people in the data recovery industry who agree with it not being possible.
You can't just say something is possible with no proof. Do I really need to say this here?
Still, if you can show me that it can be done, I'm all ears.
Please keep in mind that the Gutmann article was written a decade ago about hard drive tech that was already old at that time.

The ONLY difference between our two methods is that mine takes 3 times longer. Until you can provide real EVIDENCE, not hearsay, of someone trying this, what equipment and methodology they used then I'll have to stick with my way.
Wait a minute! You are making a bold statement about something that has never been shown to work and _I'm_ the one that has to show some evidence? You can't show a single case, anywhere, where someone has recovered overwritten data to a useable degree and all of a sudden I'm the one with the burden of proof? Come on - now I know you're yanking my chain.

Still, believe what you want to believe. I'm just amused to see the parallels between this how other people's beliefs are handled in this forum.
As soon as someone shows me that they have recovered data in this way, I will shut my big fat mouth. Gladly, too, as my job just got a hell of a lot easier.

[egslim]

For what it's worth, until someone shows evidence to the contrary you've convinced me.
It makes a lot more sense. First, it is my impression the government is generally technologically behind private companies and definately seldom ahead. Second, with multiple GB discs it would be a lot of work to repair every single bit after it has been overwritten once, one at a time. I imagine the traces of previous information would be lost in the noise due to QM effects.

[Gromit]

/\ Hooray! ;)
Mind you, I'm just happy that people will actually do some research on this topic and look for evidence. When I first decided to post in this thread I thought it was a rather ironic topic, given the lack of strong evidence and yet the intense strength of the belief in the computer industry.
I get quite a lot of flak from logical "computer nerd" people whom I discuss this with on other forums. It seems to be a belief that is ingrained and very popular. Fair enough too - I mean, it sounds pretty reasonable on the outset and hard drives are almost magical in how they function anyway.
I just find it interesting to discuss it with logical thinkers when all they really have to back up their claims is a belief, rather than any evidence whatsoever. I'm not singling anyone out with this - I shared the belief myself until I actually spent the time to do some research.
During that research I remember seeing an ad on the web from a data recovery franchise with offices all over the globe. They made mention of overwritten data, and even had a picture of an electron microscope on their web page. I sent them an official email from my work email account requesting further information. After a couple of replies they finally admitted that they used EnCase software to recover deleted files and had no ability to recover overwritten data. Good old false advertising.

Still, I'm hoping that some day in the very near future I will be proven wrong.
My work will become a lot more interesting if that's the case. The problem with this, though, is that computer tech moves so quickly if you don't get a solution within a decade you can almost forget about it. I've not done any research, but I imagine that hard drive tech is soon to change again to allow for greater capacity in the same form factor. How much further can we push current designs? 600GB? 1TB?
Hmm, I smell more research coming on.

[egslim]

As Penn & Teller said, everyone has their gree-gree's - things they believe without evidence. And those who think they cannot be fooled themselves are most vulnerable.
Of course I'll change my opinion back again as soon as someone provides some evidence. But just like with the Million dollar Challenge, I won't hold my breath...