I guess two separate hard drives from different manufacturers, stored with care in two different places.
As the new technology comes along to replace HDDs, transfer the data onto that as well.
I can remember a friend of mine storing stuff on video disks, and then a few years later couldn't retrieve it, a bit like Betamax etc.

Bill

I can see a hell of a lot of personal photos / images which will one day be historically important / interesting to someone just lost and forgotten…

I regularly handle 450 year old rag paper documents – still as readable as the day they were written – catch is their information density is somewhat low but would 450 old documents describing early computer viruses specific to long dead O/S be of any interest.

Earlier this year I went to a morning meeting held by the Gloucestershire County Archives for local history groups. It was really a sales talk but subject of digital archiving was included. It appears to be a very, very serious problem with formats changing every few years and unstable storage devices.

If you want to take this further I would suggest you contact your local county archives for advice. They may even welcome the query as a change from looking at parish records.

Today we can read day to day letters written by the Romans on clay tablets. In 2000 years time will anyone be able to read anything on an i-pad?

JA

+1 for cloud storage

A separate issue for me as a photographer is the lack of a permanent record of digital images. On the trashier channels of cable TV one sees Hitler, Churchill etc in otherwise insignificant photos. With digital these will be deleted.

How will historians of the future manage?

It's an interesting and complex topic. I used to work on the multimillion £ Digital Preservation project at the British Library. As part of that we built a system that automatically copied each digital file to 3 remote storage sites, it automatically checked each file on a regular basis to ensure it hadn't changed and if it had changed recovered it from a remote site. Files on hard drives can spontaneously have bits changed and of course there can be bugs and physical problems like fires. We archived digitised books and newspapers as well as websites and other 'born digital' files such as the huge master files for Ordnance Survey maps. When one disk started to fail it was swapped for a new one and data automatically copied over. I am sure that project has evolved since then.

Files stored on older CDs can be unreadable as the surface degrades, generally they are poor way to store data long term.

90% of the data collected during the Apollo program is now unreadable – the data format was never recorded but the data itself is fine.

For files created with a specific program you need to store the program with the file or instructions on how to read the data – no point in having a well preserved file if you have no idea how to read it.

The BBC had a project in 1981 called the Doomesday Project to record census data and general information about life in Britain – the data was stored on Laser Disks. I understand there is only a single, working Laser Disk player in existence now that is capable of reading this and this is used as a classic example if digital obsolescence. So it can be a hardware problem as well as a software one.

Data loss isn't always accidental, governments can try to change data – look how Trotsky was erased from photos after he fell out of favour in the Soviet Union. So a good preservation strategy has to cope with deliberate data tampering as well.

A great example of a failure to think the problem through. When laser disks became obsolescent there should have been a plan in place to transfer the data to the next mass storage technology – preferably in the cloud. As always, this isn't a technology problem, it's a process and system failure. I've recently been reading up on Bletchley Park again, and what comes through from a couple of the better books is that, whilst the technology was key, it would not have worked without a carefully constructed and maintained process to industrialise codebreaking.

The Wikipedia article on the BBC Doomesday Project makes interesting reading, not just to show the issues surrounding the preservation and access to this information, but also the copyright nightmare associated with trying to do so.

BBC Domesday Project on Wikipedia

Not a big fan of 'cloud' storage. My previous employer engaged an outside organisation to archive all their records. For some years now they have been progressively converting all their paper filing to digital archives in-house but the task was getting too big. As a consulting engineering firm they had to make sure design information on each project was available for the life of the project in case of any future liability claims. The statue of limitations in most Australian States can be extended where civil suits pertain to engineering designs. Seems like their highly reputable international data storage contractor lost a big chunk of stuff pertaining to some quite notable projects from the last 10-15 years. They are not happy, and neither are their insurers as they now have no evidence to build a defence against any potential claims.

A fair cop.

Spelt it wrong three times but only corrected it twice.

But as Meatloaf sang 'Two out of three ain't bad'

See:

http://www.domesday86.com/

This is a project to create interfaces so that the original data (ported to a SCSI drive) can be viewed as was originally intended – on a BBC Master series microcomputer.

Plenty of valid reasons for owning a large collection of computer viruses even if you aren't a super-criminal!

I agree with Danny's total ban, and have reinforced it in practice by configuring computers so they can't read or write media, and are fully encrypted, plus keeping virus and malware protections active, and with all the data in a physically secure computer room, not on local hard-drives etc. This is because people (including me) are unreliable.

Education is essential, but often ineffective. Unless someone has a particular interest in computer security, most people in my experience fail to pay attention or can't maintain their guard over long periods. There's another group who don't believe any of it applies to them: they see no harm in writing their password on a label under the keyboard, or copying all their customers bank details on a memory stick and then leaving it in a pub.

The level of protection required is related to the value of the data being protected and the risks to it. For example, the CIA has to manage the threat of well-funded professional attacks mounted by any of several foreign intelligence services, including Allies. Elderly gentlemen using XP only to browse this forum, never spending money online, viewing porn, or sharing private details willy-nilly on social media can be more relaxed about computer security. But no-one is immune!

Security gets really interesting when balancing usability against security needs. Security experts are only happy after the entire system has been disconnected, shredded and incinerated. Not possible, so real computer systems doing real work, always have loop-holes and vulnerabilities. It's the job of the security team to manage the risks, a particularly hard job when most of us don't understand security, are naive about the risks, get lazy, make stupid mistakes or take bribes.

Corporates often use as many as three or four different AVM products in parallel as a way of reducing virus risks: one on the internet gateway, another on all the workstations, and a third on the servers. Developers and system administrators loading new software may add a fourth. It's necessary to update all of them regularly; ideas about this vary, but weekly isn't enough! The most at risk systems will have someone monitoring virus threats continually, perhaps reacting 24×7 to dangerous issues by applying patches or by disconnecting services. Also important to keep a careful eye on AVM products because they aren't consistently effective over time. Untested loyalty to Brand-X is unwise because AVM products rely on an engine and a database of signatures. Both can get out of date, or – in the worst case – be unable to adapt to new threats without major rework. They're not like quality physical tools that stay reliable for decades.

Dave