Secure file deletion

[Goatboy]

I don't know exactly what they look for, but basically at that level they can tell, for example, that a particular bit was a 0 to begin with, then a 1, then a 0 again, 1, , etc. Sorry I can't explain it better. Google it, perhaps?

[sanddbox]

I don't know exactly what they look for, but basically at that level they can tell, for example, that a particular bit was a 0 to begin with, then a 1, then a 0 again, 1, , etc. Sorry I can't explain it better. Google it, perhaps?

*has seizure from comma overload*
*goes into a comma coma*

[Goatboy]

Hey, all of those commas, with the exception of the double comma typo, were properly placed. However, understandably, I can see how someone like you may, in certain circumstances, become overloaded by the number, frequency, and density of commas in a sentence, and, in turn, slip, oh, so tragically, into a coma.

That enough for ya?

[106627bg]

Sorry for my laziness, here's a good wikipedia article that explains it fairly well. To summarize:

The need for multiple overwrites is not proven meaning that no organization (public or otherwise) will admit to being able to recover overwritten files. On the other hand, the old polarization can linger on the memory of a hard drive. Theoretically it should be possible to recover an entire file or even an entire disk with some methods of recovery.

[sanddbox]

Hey, all of those commas, with the exception of the double comma typo, were properly placed. However, understandably, I can see how someone like you may, in certain circumstances, become overloaded by the number, frequency, and density of commas in a sentence, and, in turn, slip, oh, so tragically, into a coma.

That enough for ya?

....................../´¯/)
....................,/¯../
.................../..../
............./´¯/'...'/´¯¯`·¸
........../'/.../..../......./¨¯\
........('(...´...´.... ¯~/'...')
.........\.................'...../
..........''...\.......... _.·´
............\..............(
..............\.............\...

[jovial01]

What do you all think is better:
Deleting a file then trying to overwrite the memory repeatedly to obscure the evidence of it being there in the first place?
OR
Digitally shredding the file making it mathematically/computationally infeasible to put back together again?

A combination of the two seems better, but if I had to choose, I would choose the last option. Any thoughts?

Also, what do you all think is the best method for shredding files, do you trust 3rd party software? How about writing your own algorithms? Ideas?

[Goatboy]

I'm not sure what you mean by digitally shredding. Did you think that they "shread" the file into little places and put them all over the drive? That would simply be ridiculous. Digital shredding means erasing the file, as you mentioned in your first option.

[msbachman]

What do you all think is better:

Digitally shredding the file making it mathematically/computationally infeasible to put back together again?

Going off what goatboy said:


That's not what shredding is, though. Shredding is just overwriting existent data with other data, often repeatedly. As long as it's actually being overwritten, one pass is sufficient to make it unrecoverable. Don't buy into the nonsense that you need to make multiple passes, or follow some stupid DoD standard of 35 passes or whatever. A single pass is enough.

You can, theoretically, recover data after that, but it's not a simple matter anymore. We're talking, at that point, about electron microscopy to see vague, ill-defined, skips in overlaps where a bit might be readable from what it originally was. Even if it can be done (read up on literature and try to find where it has been done successfully on a significant amount of data...GOOD LUCK!), at that point it would be cost prohibitive for anything but the most dire of circumstances.

Related: check this out http://www.anti-forensics.com/disk-wiping-one-pass-is-enough

Also, if you're truly concerned about removing all traces of an item, you'd do well to look into full-disk encryption and skirt this issue altogether.

[insomaniacal]

You can, theoretically, recover data after that, but it's not a simple matter anymore. We're talking, at that point, about electron microscopy to see vague, ill-defined, skips in overlaps where a bit might be readable from what it originally was.

Now just imagine trying to build a 10kb text file back by manually observing these things, and you'll realize that no one will go this far unless you've managed to delete the codes to every nuklear weapon in the world.

[jovial01]

thanks for your comments.
Yeah I misused terms. I guess I was wondering how useful it would be to obfuscate the file before you actually delete it, but I guess disk encryption covers this. Also it seems simpler just to overwrite repeatedly. For some reason I had this idea in my head of maybe getting a binary representation of the file and splitting it at random points, recursively, recombining, and then doing it again, much like shuffling cards?
Im just thinking here, I realise its probably a waste of time seeing as people have mentioned just overwriting repeatedly only a few times would be sufficient...
Just thinking out loud.. Any comments welcome.