Hole in the Umbrella: Backup 3

Bottom-line, on top: Don’t use Backup 3. Use SuperDuper or Retrospect.

Backup 3 is easily the worst piece of software Apple is shipping today.

From an evil genius perspective, Backup 3 is a masterstroke. I can’t think of a better way to subtlety inflict maximum time-released damage simultaneously across so many different axes: Apple, .Mac, the backup software market, and its hapless users.

Like a Hitchcockian horror movie, Backup 3’s story features a superb set-up. It comes from Apple. You pay for it (indirectly, with a .Mac subscription). It’s elegant, with a beautiful icon. The website has high production values. Indie developers are offering extra Quick Picks. The app feels competent. It feels like something you can trust.

Of course, you can’t.

Backup software is hard and thankless. A sizable minority has bad ideas about how backup does/should work, and what their real needs are.

Backup software is an organization’s software-quality acid test. For Apple’s software spread, I would place it only after the file system and Core Data in the ranking of software that absolutely, positively must-work. It requires an engineering backbone of steel, sometimes even saying “no” to data-compromising requests from users, upper-management and marketing.

Backup 3 fails exactly where backup software is not permitted to fail: restoring data. Hardened sysadmins know to initially and continually test data restoration. What’s the point of backing up if you can’t get your data out later?

And this is why Backup 3 is so damaging: its target demographic is exactly the type of people who don’t know — or lack the hardware resources — to test restorations. Backup 3’s high production values can lull even those who should know better into a false sense of security.

Unfortunately, the harm doesn’t end just in the black eye for Apple’s software quality-control, the valuation of .Mac, or the users who will lose data. Backup 3 also hurts the market for backup software that actually works.

The Mac backup market is already in bad shape. You have the old-tymer Retrospect, which has the best engine but suffers from an atrocious user interface and inability to recycle space on a backup volume. You also have the promising upstart SuperDuper, which looks great in most every way except its lack of “historical rollback” (unlike David, I consider temporal versioning critical).

Today, there isn’t a single backup program I can recommend. Eventually, either Retrospect will get easier and work well with HDD-based backup, SuperDuper will gain temporal versioning or Apple will fix Backup. Tragically, any of these scenarios are currently probably more than a year away, if that soon.

Update: I don’t know how I get off so lucky, where even my hate mail is lucid and well-argued. It would be so much easier to ignore if it was incoherently frothy. Adrian Milliner writes:

You said:

“…Backup 3 also hurts the market for backup software that actually works.”

I’d submit that the your comments are more damaging than Backup 3 itself. Backup 3 has its flaws - all software does - but it only fails for some people some of the time. You damning comments are far more widespread.

In a similar vein to your piece:

RSS bloggers, like real journalists, have a responsibility to be accurate, balanced, and be able to back up (no pun intended) what they claim. OK, so you don’t directly say that Backup 3 is incapable of restoring your data, but you clearly give that impression.

Just my £0.0112

I’m an admitted software perfectionist, but recognize my predilection is nonproductive in most cases (otherwise I’d flame the Finder every other day).

Backup software is a quality hair-trigger for me. There’s zero tolerance here, a binary state. It’s not a matter of design opinion, like the sheen on iTunes — it’s an absolute fact: either it works or it doesn’t. The only leeway here is that “working” software is allowed to fail during backup, if it does so with notification (for example, running out of space). Hence “only fails for some people some of the time” is, in my estimation, the same as “completely broken”.

I did give an impression, so I’ll be crystal: I don’t trust Backup 3 to restore data. I have not personally verified Backup 3’s restoration failures, the extent of my evidence is Erik’s and Michael’s writings. Both these guys are Apple-supporting software professionals whose opinions I respect, so when they talk software, I listen.

My bias: the reason I have not personally verified Backup 3’s restoration failures is that I don’t use it. Even though I dislike Retrospect and want to move off it, when Backup 3 shipped I privately laid odds it would be untrustworthy. Hence I followed Erik’s move to Backup 3’s intently, with a mixture of morbid fascination. Would he hit the iceberg I predict lay in wait?

Speaking to the blogger responsibility point, I did not blog my prediction or even mention it to Erik. I think it’s clear I’m an Apple fan, and all I had was a hunch. I wanted to be wrong. I didn’t even post here once Erik had his restoration meltdown, since it was only one data point. It was Michael’s posting that pushed me from private seething to public condemnation, and explains the vitriol of this posting.

Apple is good at delivering what we want. It’s our responsibility to send a clear message when they fall dangerously short, so they can fix it.

Update: Dave Nanian of SuperDuper talks about temporal versioning. He sticks to his guns and makes good points. Hey Dave, let’s make a deal. I’ll concede most of your points, if you drop the exclamation mark from the product’s name. Act now! Supplies are limited!

Update: Josh Carter offers tips. The loss of resource forks appears to be related to manual restores.

Saturday, October 29, 2005
12:00 AM