So goes the story: In 1837, Thomas Carlyle finished his 800-page opus, The French Revolution: A History, and offered his friend, the philosopher John Stuart Mill, a first read of the single finished draft. Mill eased into the text one evening in his study, by the fire. He read a few pages, became sleepy (critics are still divided on whether the text or the time of day was responsible), and turned in for the night. He left the book in the study.
Mill’s maid found it there, somehow mistook the pages for waste paper, and used them to light a fire.
Data loss is not only a danger of the computer age. I wonder what Mill said to Carlyle the next morning. Sheepish doesn’t quite cover it. At any rate, Carlyle rewrote the entire book. As did T.E. Lawrence, when the first draft of Seven Pillars of Wisdom was stolen with his briefcase while he was changing trains at Reading. As did, and still do, many other writers who lose work by way of the vagaries of daily life.
If you don’t mind this type of creative process, what might be called editing by karma, then by all means write your book and leave the only draft on a train (or by the fire). But if you prefer more security, more control of the means of production, you need a good system for backup. Here’s mine.
First, write by hand on paper whenever possible. Transfer your text to the computer, every day if you can manage it. That’s one backup. Specify in your word processor settings that all files are automatically backed up. That’s two. But what happens if your computer crashes, or is infected with a virus? Besides, your hard disk is likely designed with a ten-year lifetime, and there’s always the unintended ‘failure rate’ of components, which in practice means your data is not secure (on a single disk).
Backup to as many hard disks as you can find. You likely have some unused disk space at your ISP (the people who connect you to the Internet: normally the phone or cable company). You may have a home network, or a website. Use these as remote backup disks. Don’t rely on floppy disks, or even CDs: use multiple, separate hard disks. Make the backups automatic, (e.g. every night), and place an icon on your desktop for instant backup after you finish a writing session. One click and you’re done.
But how to setup all this wizardry? I can’t speak for Windows users (in other words, almost everybody); but on Linux the process is easy. I use a program called rsync, which I believe can be made to work under Windows if you know geek-speak (in this case, how to compile programs). And the web is a big place: synchronization software can be found all over: like here, for example. Whatever you use, take the time to get to know how the software works, what it will do, and what it will not do.
My setup works like this:
Every night at 11pm, and at noon every day, (using the Linux task scheduler, called cron), or whenever I click on the desktop shortcut, rsync goes to work. This is what it does:
Copies all local changes from my writing files to a private folder on my website.
Copies the same changes through the wall and onto the networked computer in the other room.
Copies the changes to a second hard disk on the networked computer.
Copies the changes to a backup folder on my local hard disk.
Copies the changes to a second private folder on my website.
Emails me a friendly notice that the backup has been completed (with details).
All the commands (except the last one) have a similar format, and I just list them in a script file. Each is a variant of this:
rsync -varuzP /source/directory/ /destination/directory/
The ‘varuzP’ settings are switches for rsync: verbose, archive, recursive, update, and partial. If the backup requires that I login to another computer (e.g. my website), the command looks like this:
rsync -e ssh -varuzP /source/directory/ myuserid@myremoteserver:/destination/directory/
The ‘-e ssh’ setting specifies that rsync use the OpenSSH secure shell protocol, which I have setup for remote login without a password. This allows me to put the backup command into a script and run it automatically (and was the most difficult part of the backup to create, requiring some research into the arcane world of online security). Windows users can use PuTTY for ssh logins.
This all sounds a bit extreme, doesn’t it? But if my house burns down, and all I have are floppies and a single backup on one computer, I lose everything. If I come home at the end of the day to find a smoking ruin with my wife and kids huddled in blankets on the front lawn, you can bet that my writing is one of the first three things I’m going to ask about. (The other two are my library and my woodworking tools. My wife will dependably be concerned about all the important, family-oriented things, so I’ll leave that to her. Self-orientation is my area.)
With my current backup arrangement, I’d have to be extraordinarily unlucky to lose my data: house burns down, website is corrupted, and remote private folders are trashed. And these various catastrophes would have to happen all at the same time. Not bloody likely.
Notice that the backup is circular: it starts by copying files to one disk (the remote website) and finishes on the same disk (though in another folder). It moves from one disk to another, cycling back to the first, so that in the unlikely event of a power failure in the midst of a backup, at least one disk is likely to be up to date.
Once a year, I backup onto CDs. I’ve never used these backups, and I never expect to. The data is almost immediately out of date. Besides, finding the right CD is a chore.
All told, from learning the software to setting up the automatic logins required to get files uploaded to my remote backup locations, I’ve probably spent twelve hours on my backup scheme. This seems like a long time. It’s equivalent to perhaps a week’s writing. But in the event of data failure, my strategy could save a great deal more. I don’t typically print out my work until I have a hundred or more pages, so at the very least I could be saving that much.
My backup system is rudimentary in comparison with professional data archiving. I don’t use compression (I don’t trust it), multiple date-stamped versions, or any of the many tricks that are available. I just copy the files. But my system works, and I no longer worry about data loss.
Backing up has become part of my writing process; an important part. It’s not just for geeks.
