Skip to main contentSkip to navigationSkip to navigation
Brian Fitzpatrick
Brian Fitzpatrick is leading Google's drive to make the export of data a one-click process. Photograph: Linda Nylind
Brian Fitzpatrick is leading Google's drive to make the export of data a one-click process. Photograph: Linda Nylind

Google's plan to free your information

This article is more than 15 years old
As head of its 'Data Liberation Front' Brian Fitzpatrick's role is to make it easier to export your files from Google's servers

For years, the aim of pretty much every technology company has been to make a product that people can't give up using, and in case something better comes along from a rival, make sure that they can't get their stuff – whether it be data, software or hardware – to work easily with the newcomer's platform. On this rested the success of the compact cassette v the 8-track, VHS v Betamax, Iomega's Zip v other backup systems, and most recently Blu-ray v HD DVD.

Amidst which, Brian Fitzpatrick's role at Google sounds, at the very least, contrary. He runs its self-styled (and half-jokingly named) "Data Liberation Front" in the Chicago offices, and his aim is to make it easier – one button is the ideal – to export your data from Google's servers onto a storage format of your choice – whether that's your own web server, your computer, or the comfort of your backup drive that you keep locked away in a fireproof cupboard after using it every night.

The Data Liberation Front – the name's a jokey reference to the Judean People's Front, the would-be terrorist group in Monty Python's Life of Brian that never quite gets its act together and spends most of its time bickering – is actually a good thing for Google's customers, Fitzpatrick argues, because it means that lock-in element can't be applied to your data.

"Think of it like you were renting a house," says Fitzpatrick. "If you decided to move out and the landlord came and told you that you couldn't take your furniture or your clothes or your family photos, you wouldn't be pleased, would you?" His point being that Google wants to give you that comfortable feeling that if you need to export your data then you can.

In the click of time

It's already been achieved for Blogger, the free blogging platform the company bought. There is a one-click export (to the Atom format) which preserves not only posts but also comments. (An export to RSS, which is also available, only preserves the blog posts.) Google Notebook, which has been "end of lifed" (read: killed off), has had export functionality added to it. Fitzpatrick notes all sorts of Google products that have got export functionality: Google Docs, iGoogle, and various other Google products. (And, inevitably, you can follow it on Twitter at twitter.com/dataliberation — which might make you ponder how easy it is, by contrast, to get your tweets out of Twitter.)

And next, he says – though dates aren'tgi – there'll be an "export" button for Google Sites (in HTML), as well as a "mass export" from Google Docs, for those who want to export a lot of data at once.

You can see the clever sales logic. Many people fret that with cloud computing you can't walk up to any location – still less a specific machine – and say: "My data is in here." Such distributed services mean your data might be on five continents at the same time.

Thus people, and companies, get uncomfortable about trusting a cloud service, because they don't know where it is, and so can't be sure it's really safe. For Google to say "we can easily import your data" isn't more of a claim than others are already making.

But if it then says "exporting your data is one-button easy", it actually has a selling point. True, it looks perverse to those accustomed to the lock-in mentality of previous commercial battles. But it may be the right approach for the web. It's classically, Google-typically, counter-intuitive.

Fitzpatrick studied Latin and Greek ("and ceramics") at university, then went to work for OnShore, a small networking company based in Chicago. There he got interested in fixing a problem with an open source database driver, and was then encouraged to submit the change to its authors. Which led to working on Subversion, a version control system widely used by teams of programmers who need to co-ordinate different versions of programs. He then went to Apple, where he worked on the consulting team that would go with every sale of its fabulously expensive WebObjects package, and then back to Subversion. (He wrote the book on it.) When Google bought the company he was working at, he was reluctant to join: he'd set down roots in Chicago. But the company was happy to let him set up an engineering department in the city (it already had a sales centre). He's also in charge of Google Affiliate Networks, an acquisition from the takeover of DoubleClick. He adds: "We believe in an open web for everyone … The web is fundamentally about openness."

Open and shut case

But there's also two other ways in which it works to Google's advantage. First, it encourages its developers not to fall behind rivals. If the price of being overtaken is that people will pick up their data and leave your application behind (which might then mean your job as the application's developer vanishes), you'll have a stronger incentive to keep going. But equally, for managers who don't want to have to support a million wilting blooms, being able to export data means that unsuccessful projects can be shut down without regrets that users will curse the company for locking away their data on its servers forever.

Compare that with the outcry that Yahoo faced when it announced it would close Geocities: efforts to save it sprouted up, and Yahoo wasn't popular. Google isn't popular for closing services – but at least Google Notebook users can get their data out.

So, export for blogs and Google docs is straightforward enough, as everyone is familiar with their formats. But how will exporting work for a completely novel idea, such as Wave, whose functionality nobody outside Google ("or inside," adds Google's PR woman, who is listening) has yet managed to describe in fewer than a thousand hand-waving words (it's something like "email and instant messaging and collaboration but with changes shown over time")? How do you export something which has a unique format?

For a moment, Fitzpatrick looks faintly alarmed. But that's not because he hasn't considered it – although Wave was developed in Sydney, his Chicago team has already been looking at what it needs to do.

"We have talked about it. It's not that difficult to represent [its data]. The question is how to represent time. Wave has the extra dimension of revisions. There are ways to represent that but nothing else really has anything that it's like. It's unique." What about Wikipedia's "diff", which shows the differences between revised versions of the same page? "That's perhaps the closest," Fitzpatrick acknowledges. The problem then is that a diff is a database representation and there isn't an agreed way to export a database. (SQL ends up being database-specific, Fitzpatrick says.)

The irony is that if Fitzpatrick succeeds, then Eric Schmidt, Google's chief executive, will probably be happy. "He keeps telling us, the way to not be evil is to not lock users in," Fitzpatrick says. "He tells us, just get the users and we'll figure out how to make money."

Comments (…)

Sign in or create your Guardian account to join the discussion

Most viewed

Most viewed