[process] Backing up your fiction like a pro

Nobody has asked this question, but I’m going to answer it anyway: “Jay, what is your data backup strategy?”

Given that I’m in the midst of drafting a novel right now, this is of immediate and daily interest to me, I’m glad you asked!

I have been using word processors since 1985. I have had all the usual sorts of failures — corrupt hard drives, video blackouts, dropped equipment, water damage. Once, in a hotel room, a picture fell off the wall and smashed my laptop where it sat on the desk. I’ve even had computers catch fire while I was using them. (Twice.) I was a PageMaker 1.0 user and QuarkXpress 1.1 user, which means I have lost cumulatively hundreds of hours of effort to unexpected application quits and corrupted files.

In other words, I never trust my computer hardware and software to remain safe, stable and accessible from one minute to the next. And since those files on that hard drive contain the words which are literally my stock in trade, I take very good care of them.

First of all, I have my Mac’s hard drive formatted into three volumes. One is /sys, for the OS and application set. I store none of my writing or personal data there, and having the separate partition for system files only allows me to occasionally perform stupid computer tricks which can be quite helpful. One is /writing, which in the days of CD burning drives I limited to 600MB so I could burn the entire volume to a single CD-R. Now I’ve raised the size to 4GB, which I can fit on a single DVD-R. I also have a third volume for everything else — photos, music, miscellaneous files, things which don’t need the backup rigor of the writing data. That way I’m not trying to juggle dozens of gigabytes of data in the high-demand version of my backup process, just the writing files.

Second, to guard against file corruption, every time I complete a novel work session or a short story draft — and sometimes work session there as well, if it’s a long piece of short fiction requiring multiple writing session — I make a copy of file at the operating system level. In other words, duplicate it rather than “Save as…” within my word processor. This creates a trail of files which may be 100 deep for a novel project. Each file captures the work at that time, which gives me a backtrail if I need to roll back a revision or writing effort to some previous state. More importantly, if my file has begun to corrupt somewhere, and several of the most recent copies are flaky, I can backtrack to the last good file.

By the same token, I use version control numbers on my files. That means when a piece of work is finished, it becomes “Story 1_0.doc”. As opposed to “Story 0_1.doc” and “Story 0_1 copy.doc” and “Story 0_1 copy 2.doc” and so forth. When I begin to revise it, the file becomes “Story 1_1.doc”, and begins to generate its own trail of duplicates.

Third, I send the copy of the most recent work session (not the master file) to myself via gmail, as well as to a trusted friend in another city. Gmail means the file is stored as an attachment on Google’s servers, as opposed to, say, emailing to myself via a POP account and Eudora, which would have the net effect of storing the file right back on my hard drive. Now if my entire computer goes sideways, and I do not have access to other backup mechanisms, I can at least manually retrieve the work by logging into gmail from any computer, or having my friend mail my file(s) back to me.

Fourth, if I am at home, I run Time Machine backups automagically. I have a 1TB desktop drive and a 500MB portable drive which I swap out daily, so both are carrying the most recent Time Machine backup.

Fifth, when I travel, I take the 500MB portable with me, and run Time Machine at least daily on the road. I also take it when I write at a coffee house or a write-in, or anywhere away from my home office setup.

Sixth, when I am away from home and do not have network access (on an airplane, for example), I also copy the file duplicate mentioned above onto a thumb drive which I store in the portable Time Machine drive’s carrying case.

That’s my second-greatest point of vulnerability, by the way. In the event of an airplane crash, for example, I would probably be forced to abandon both my computer and my backup drive to evacuate. My plan in that situation is to stuff the thumb drive in my pocket and hope I don’t have to swim for it. (And yes, I’ve thought even that through.)

Finally, every two or three weeks, I burn my entire /writing volume to a DVD-R and mail to my aunt in Tennessee. That guards against a regional catastrophe such as a massive earthquake, tsunami or volcanic eruption, ensuring that my data will still be available even if my entire city has burned down. (And yes, I’ve also thought that through.)

My greatest point of vulnerability? Forgetfulness. Friday I was writing with in a coffee house on Hawthorne, and I’d forgotten to bring my portable Time Machine volume with the thumb drive in its case. It had been out on my desk at home for a routine backup, and I’d simply not paid attention when I left. As I was wrapping up my work session on The Heart of the Beast, I went to gmail it to myself and my friend when the network simply bagged out. The file wouldn’t attach. And I didn’t have my thumb drive.

I felt as if I’d shit on my hands and couldn’t find a washroom. I literally itched. It bothered the hell out of me. What if the drive corrupted on the trip home and I couldn’t get back into the data? What if I had an automobile accident and the computer was physically damaged? I was stuck for it, so I went home. As soon as I got there, I rectified the problem, but I was extremely conscious of the problem the whole drive home.

Other things I do to protect my data:

Save my work often. I don’t use “Autosave” type features because I find them deeply intrusive, but I’ve trained myself to hit the save key command reflexively. That protects me from an abrupt application exit, though I still might face file corruption issues.

On long work sessions, stop in the middle, make a duplicate file, and send it out. Then resume writing. This also helps protect me from file corruption issues.

Never, ever, ever leave my laptop in a vulnerable location if I can possibly help it. This means an ironclad rule about never leaving it in the car, not even locked in the trunk. As it happens, I drive a convertible, so I never leave anything in the passenger compartment anyway. Even junk, like an empty box, might look valuable to a meth-head or a vandal. I could get my top cut ($2000+ to replace) for the sake of a worthless piece of cardboard. Likewise, I never lock my car for the same reason. But even in the trunk, someone with a lock jimmy tool or a crowbar could have it out; or they could just steal the whole car.

I can replace the car a whole lot easier than I can replace the computer.

Occasionally when I travel I am forced to leave my laptop in a rental car. This makes me very uncomfortable, but that’s when I just have to trust that my layered backups are sufficient.

One thing I don’t do to protect my data that others recommmend:

I don’t keep printouts of everything. I used to do that, years ago, but keeping up with printouts of my work is like the sorceror’s apprentice keeping up with the mops and buckets. I simply don’t have the space, time or paper budget to do this. Even without saving drafts and interim files as printouts, I should probably keep printouts of finished material, but even then I’d have a stack of paper at least ten thousand sheets tall.

I’m also certain there are a dozen avenues of vulnerability that I haven’t addressed. I’m no business continuity guru or data security expert. When I find them, or have them pointed out to me, I modify my backup strategy accordingly.

The words are the most valuable material thing in my life. Due to my work style and choice of tools, I keep them in an inherently unstable container — electronic files on my computer. So I do everything I can to keep them very, very safe.

8 thoughts on “[process] Backing up your fiction like a pro

  1. John R. Douglas says:

    Hey, Jay:

    Have you invested the time on things like Dropbox or Wuala or the time and money in things like Carbonite (or many other alternatives) for automated online backup/syncing of selected files or all of your computer files. I’ve not had occasion to need them (Yet!) but I have all three of these systems operating on my main computer at home. Seamless, essentially invisible online duplication of files. Not always absolutely up-to-the-minute but usually pretty close.


    John Douglas

    1. Jay says:

      Hey, John. That’s a good idea. Essentially a slightly smoother version of what I do now with Gmail, and it offers some possibilities for collaborative work. Thanks!

  2. Jaws says:

    This is generally good advice — although I would not trust a third-party vendor with unpublished drafts, that’s really a personal decision — but it neglects something that’s pretty important:

    data neutrality (third bullet point)

    Had I not had the foresight to save all of the data files supporting my master’s thesis in CSV too, instead of just the now-defunct database format from which they were drawn, I wouldn’t be able to even access the raw data.

    1. Jay says:

      That does it. You’re going into tomorrow’s Link Salad, buddy.

  3. Tim Keating says:

    I’ve had all I can stands, and I can’t stands no more.

    Scalzi posted about his backup strategy, which is equally bizarre and convoluted, some time ago, and at that time I said nothing. This time, I can’t let it go. Rather than spam up the comments on this thread, I have written a blog post debunking this backup strategy: http://www.mrtact.com/blog/2009/02/how-to-really-back-up-like-pro.html. Lest you think I am some shill redirecting you to an ad site, let me say that all the software I talk about in that post is free, and I have nothing to do with it beyond a wish to steer other writers away from these antiquated, time-consuming and frankly, dangerous practices.

Comments are closed.