Follow Schofield's Three Laws of Computing and avoid disasters

These "Laws", developed over three decades, may highlight things you're doing wrong, or help you avoid the worst screw-ups. Violate them at your peril…
Written by Jack Schofield, Contributor

The New Year is a good time to think about how you are going to approach things over the next year, and preferably the next few decades. When I do this, I use Schofield's Three Laws of Computing, and you may find them useful. They're not really Laws and they're not new -- the first dates back to the 1980s -- and I have written about them before in the Guardian, but not here.

Schofield's First Law of Computing states that you should never put data into a program unless you can see exactly how to get it out.

This law was first framed when I had data on 8-inch CP/M floppy disks, 5.25-inch Apple and Atari diskettes, and various other media. Today it would obviously include cloud storage.

Many people spend a lot of time worrying about which PC, tablet or phone to buy, when they might only use it for 18 months to 3 years. Hardware is transitory, and rapidly becomes either scrap or landfill. It's much more important to think about software -- which often lasts for decades -- and data, which might well outlive you. Data is forever.

This law is particularly important when data costs much more than either hardware or software. Just estimate how much it would cost to retype or recreate data rather than import it.

The most monumental errors arise from ignoring the First Law. One example was the BBC's Domesday discs, which were created to run on 8-bit Acorn BBC micros in UK schools and public libraries. This was supposed to be a new version of the 1,000-year-old Domesday Book, but both the computer and the video discs rapidly became obsolete, making the data inaccessible. It was rescued, as I wrote here, but it would have been much more sensible to use a commodity hardware platform and open (or at least ubiquitous) file formats.

Schofield's Second Law of Computing states that data doesn't really exist unless you have at least two copies of it.

If you only have one copy of something, it's a bit like Schrödinger's cat. Is it alive or dead? You have to look in the box to find out. In the same way, you'd have to load your single copy of a file or database to see if it was "live" data or corrupted (dead) or had simply disappeared. (Perhaps you deleted it by mistake or forgot where you stored it.)

Ideally, you should have at least three copies of everything, preferably on different media. It is a good idea to store one copy in the cloud, as then you have data "off premise" -- buildings have been known to flood or burn down -- as long as it's not your only copy. Having three copies means you can do file comparisons and therefore check if one of them has been corrupted.

Schofield's Third Law of Computing states that the easier it is for you to access your data, the easier it is for someone else to access your data.

In the days when everything was on an inaccessible mainframe and backed up on tape, data rarely went walkies. Nowadays, important business data can be stolen using a tiny SD card or USB memory stick, or lost when a smartphone is left in the back of a cab, or a laptop is left on a train. Data stored in the cloud may be vulnerable to hacking, especially if the hacker has physical access to a hard drive where the user has stored passwords for easy access.

Samsung smartphone
You get a nice new Samsung smartphone, but do you know where your data lives, or the file formats used to store it? Photo credit: Samsung

Having access to someone's web-based email account can be as good as having their smartphone or laptop, and sometimes better. People often email themselves important files, or store them in a related Gdrive or SkyDrive. Further, some simple searching will usually fish out loads of plain text passwords. If not, you can easily get a password reset, unless the account uses two-factor security.

The proliferation of poorly-protected, unencrypted BYOD hardware and cloud services may mean there is now less chance of losing data under the Second Law, but more chance of losing it under the Third Law.

The First Law reflects the 1980s when there was a huge variety of hardware and software, and the Second Law reflects the 1990s, when people relied on cheap hard drives that were rarely backed up properly. The need for the Third Law came with the proliferation of devices and web-based services in the 2010s. However, I don't yet see the need for a Fourth Law -- though I'm open to suggestions. The First Law already covers the growth of incompatible cloud services that work like lobster pots, or the Hotel California.

In a better world, I'd be able to advise people not to put data into a cloud service unless they could export it to a different cloud service without downloading it, possibly reformatting it, then re-uploading it. Until you can do that, the idea of having everything in the cloud is, frankly, insane.

Editorial standards