KEEPING IT SAFE

Dick Pountain/18 June 2007/14:06/Idealog 155

This month's topic - backup/archiving - seems to have just chosen itself by cropping up just about wherever I look at the moment. Jon Honeyball recently warned us about the legal need to keep untamperable archives, and I've just started writing another book so I'm in that state of "normal paranoia" whereby I need to carry a copy of it whenever I leave the house unattended (a USB key works fine). Meanwhile on Cix people are scrapping over whether Acronis is a better image backup utility than Drive Snapshot or whether they're both too dangerously fallible to rely on for system restore.

I personally gave up on image backups once my hard drives got too big to store an image on a single CD. Instead I mirror my hard drive to a little Lacie external USB hard drive, copy all my important data to a 1Gig USB key too, and resign myself to to reinstalling everything in the event of drive failure - I do keep a folder called Program Zips with the install sets for almost all the software I use for this very purpose. I'm sure Jon H's advice would be to take virtual machine snapshots of my working Windows to an external drive, and some day soon I might get around to that. It still seems to me not far short of criminal that we're still left flailing around with such a mish mash of ad hoc backup solutions, and that it's never struck Microsoft as sufficiently important to make backup a seamless, invisible, routine function of the Windows OS. Backup becomes more and more important thanks to the spread of digital photography, movie and music making which leads even the most casual of amateurs to rapidly accumulate hundreds of gigabytes of data that mean a lot to them. In my own case my photographs can be stored safely(ish) on Flickr, so it's books and columns that I really worry about, and they all still comfortably fit on a 1GB USB key.

But all this thinking about archival storage set me wondering about preservation of data in a much wider sense. Many columns back, eight years ago in issue 56 to be precise, I touched on the subject of long-term preservation of digital data, and mentioned the National Archives in Washington which has to maintain a variable-rotation-speed Memovox phonograph because the US Army stored 70,000 recordings in that format during WW11. There's now a generation of kids who've never even seen a floppy disk let alone a Memovox, and the music business is abandoning solid media entirely in favour of intangible downloads. Just who is going to archive all this stuff for posterity? The horrific pillage of the Baghdad museums during the Iraq war should warn us not to take the safety of the past for granted.

One problem is who will pay for the labour involved in digitising all our cultural artefacts, and a second, technical, problem is how can we ensure the resulting data can be preserved and read in the future. The first is at least being talked about, with Google offering to digitise the world's main libraries (while it has the money) and many of these museums and libraries doing sterling work on their own collections. The second problem is much harder. The US Library of Congress's National Digital Information Infrastructure and Preservation Program has been voted $100million to look at the problems of digital preservation (see http://www.digitalpreservation.gov/) but you'll not find any radical technical fixes there, just talk about the painstaking processes of refreshing, migration and replication: copying data from one soon-to-become-obsolete medium to its successor (say from CD to DVD). That assumes there'll always be funds and people available, to say nothing of electricity, computers and law-and-order. Given the variety of threats we currently face I wouldn't bet on all these conditions pertaining much beyond 20 years...

My own modest proposal would be to decentralise the task, at least for music and the spoken word. Of the current digital media my favourite bet for physical longevity would oddly enough be the commercial CD which employs mechanical pits pressed into plastic rather than perishable organic dyes: if you can stop the metal layer oxidising it ought to last a good while. Solid-state Flash memory as in an iPod or USB key sounds tempting - no moving parts - but it only retains its contents without power for around 10 years. So, persuade lots of people to make "time capsules" (remember those?) out of tough metal cylinders filled with a pile of CDs of the greatest music, spoken word and talking books. Flush these cylinders with nitrogen before sealing them to keep out oxygen, then bury them. How about playing the disks in future? A converted DiskMan-style portable CD player with its lid covered in solar cells, that starts playing as soon light falls on it - that goes into the cylinder too. Designing something like that ought be a weekend's work for some hardware nut, and then we need to start a campaign to knock these devices out cheaply to people who want to participate. My guess is there are warehouses full of DiskMen rotting somewhere, if only the accountants can be persuaded to let them go...