Owning our data

Posted on Jul 28, 2008

Danny O’Brien’s piece about living on the edge set me thinking about data that I have stored out there, in ’the cloud’. For the purposes of this exercise data that is stored with my employer because they choose to have it that way (bug reports, expense claims, …) is ignored, even though I do cache or backup some of that locally (expense reports mostly).

  • Flickr has a couple of hundred of my images, but they are backed up at home. I only use Flickr as a way of publishing pictures, so losing access to them wouldn’t be bad. Arguable more valuable are any comments or group postings that I’ve made or received. These aren’t backed up anywhere other than Flickr.

  • last.fm knows about a few thousand music tracks that I’ve played over the last five years. It would be annoying to lose this, but not serious.

  • Tripit and Dopplr both have information about my recent and near future trips. All of this data is replicated elsewhere, either in its’ original (less convenient) form or in my calendar.

  • Google Calendar holds both my own and some family appointments in four calendars. This is my most used web application, and it’s not one to which I am ever likely to have the source, though I can get my data out easily (I already do this for backup purposes). Having spent an hour looking for an open source alternative, Zimbra comes closest. Part of me still pines for DateBk5.

  • Google Reader knows my list of feeds and how they are categorised. It wouldn’t be terrible to lose this - I’ve used other feed readers in parallel with Google on occasion and could survive.

  • Twitter, identi.ca, Tumblr, delicous and others also have a small chunk of my data, but nothing that I couldn’t live without (or easily replicate, in the fine example of identi.ca).

In almost all of these instances it’s all about convenience. Calendaring is the notable exception, as the others can die and I will hardly notice. Scanning around the alternatives suggests that producing good centralised web applications is hard. Producing good distributed web applications is unlikely to be simpler.

When these good distributed (web) applications have been developed, getting them deployed properly will not be simple. The most significant attraction to me of hosted services like Google Apps is that I don’t have to keep my copies of Postfix, dovecot, spamassassin, etc. patched, up to date and perfect. The Zimbra desktop edition is an interesting answer to this, but it’s not necessarily to everyone’s taste.

Oh, you’ll note that this weblog migrated back from Blogger to my own machine - I’m trying to find the edge.