I worked on a DEC PDP 11/34, probably in a previous life if I think about it.

Every other Sunday we, the two programmers, had to go in for seven and a half hours overtime. That consisted of backup to a TU11 tape drive of the 2 RK07 disc drives, full format of the 28Mb drives, then restore it all from tape. There was no 'DEFRAG' back then !

( DEC RK07 - http://computer-refuge.org/dec-pics/rk07.html )
( Tech info, if you must - http://gunkies.org/wiki/RK07 )

You either did that, or watch the entire system grind to a halt if you left it longer than 2 weeks.

So, it needed two people for safety reasons. There was a big Halon gas cannister in the computer room, nobody was allowed to enter without someone outside who could raise the alarm if the cannister went off.

It took seven and a half hours to write the tapes, format the drives and restore the data, then take the tapes off-site as that session's full backup. You didn't take lunch break, the boss considered it was one long break apart from the occasional 2 minutes to swap a tape.

Most of the time we were sat reading in the programmers room. Not much else you could do. This was pre PC, pre Internet days. As long as you were ready to change the tape every 27 minutes, or whatever it took, you could do no more.

The drives had a special lid that you had to engage to get them out of the cabinet. A 'sealed' drive could be taken off site for safe storage, if we were on 'full' backup. The idea being that we could backup drive to drive and swap the physical discs over.

BUT, if you dropped the disc, there was a little blue indicator to show it was 'damaged beyond repair'. If, as sometimes happened, we were replacing the entire disc unit, the DEC engineer would be called to deal with the old disc unit. He would simply drop it, see that the indicator was blue, and toss it in a skip...!

Dafter still - the indicator was in the lid - not on the drive itself. You could, in theory, have dropped a disc, then swapped the lids over when you installed the now 'dead' platter, call an engineer out, have him replace it FOC and just tell the boss it had failed.

So much for security back then.

(In later years I managed to reverse a 1200/75 modem, supplied for diagnostics, and use a dumb terminal to access the new fangled bulletin boards - but that's another story...)

I've had a few employers who think that travel time shouldn't be counted as work.

I tend to think that everything about the way we live nowadays is fully recorded and will be available to future historians.

Your post illustrates just how wrong that thought is. Historians will think we make our living on reality TV!

To you, me, and many others our daily work experience is trivia. Historians will make a living out of it.

My info is several decades old. I used to know someone who made colloidal magnetic suspensions for researchers and forensic organisations. Of course, magnetic force microscopes weren't invented then and the amount of data recorded on disk drives was miniscule compared with today's devices.

I've only given the article a quick scan but it does not seem to mention bad blocks that have been reassigned by a device's internal firmware. These will be invisible to standard software but will be visible to specialist hardware and may contain useful information.

I'm slightly sceptical that overwritten data can be recovered from modern drives if only because we've reached the point where it takes a lot of analogue processing vodoo just to get the data which is supposed to be there.

Some of the discussion and techniques relate back as far as the (now very) old FM and MFM disks - head positioning on those was so inaccurate by today's standards that tracks were intentionally written wider than necessary and there was a lot of "overspill" of data. I can easily believe that data could be recovered from those.

Even where we're looking at research papers from 2000-2001 then they were probably based on disks with capacities less than 200G - today's drives are reaching 10-20x that (4TB drives are now available) so it's likely to be 10-20x as hard to pull overwritten data from them.

I'm not sure I'd like to state categorically that it can't be done, especially if you've got government research budgets and the ingenuity of the folks at GCHQ. I'd have loved to see them tinkering with a disk to get data off it - or even just to see them destroy their own but, sadly, although I've visited quite a few times it's not the sort of place that encourages wandering around in the hope of spotting something interesting! Even with adequate clearance (and I had that)* it's very much a "need to know" place and I didn't - all my contact was purely software issues and projects.

What I think it is safe to claim is that increases in data density will make it a harder problem simply because there will be fewer and fewer magnetic domains to keep memories of "old" data.

It is true that sector remapping is a problem for the truly security concious, not least because that data can't be overwritten by the operating system. It doesn't usually survive physical destruction though.

But all this is moot - encrypt the disk!

* And clearance would have been easy to loose if I'd been caught snooping around looking for "interesting" stuff - as billford says there are just some things you don't do in that environment.

In reply to a post by Pipexer:

.... Even the most idiotic of computer users are always talking about how to dispose their HDDs when they are done with them...

In my experience that is simply not true. As I posted earlier, with the vast majority of used drives that I see (And I see quite a lot now in my voluntary work), there has been no attempt whatsoever to even delete personal data.