I’ve been lucky enough to have been a beta testers for daytum.com, a service for collecting and communicating personal data, and I love it. As you might expect from Ryan Case and Nicholas Feltron it’s a lovely piece of interaction and graphic design. You can record and visualise all sorts of qualitative and quantitative data – personally I’m recording information about what I eat, drink, how much I sleep and communicate (emails, blog posts, talks, tweets etc.) but others record the music they listen to, how far they run, gigs they’ve been to, books they’ve read. All sorts of things.
And now you too can record and visualise whatever you want because this weekend the service came out of beta. Now here’s the thing, as much as I love the service I wish it were more, well born of the web. You see I have a few problems with daytum.
My main problem is that I can’t point to the stuff I’m recording. That graphic at the top of this post doesn’t have a URL so I can’t link to it or the underlying data; and because I can’t point to it it limits what can be done with it. If I can’t link to to, I can’t embed it elsewhere, I can’t link it to other data sources and mash it up. And that’s a problem because the only possible URI for this sort of information about me is locked away in the daytum interface. Why isn’t there a nice RESTful URL for each ‘display’. Something like:
Once everything has a URL then I want each of those resources to be made available in a variety of different representations – as JSON, RDF and ATOM for starters – that way the data can be used, not just visualised.
And finally I want to be able to use URIs to describe what I’m measuring, not just strings. I want to be able to point to stuff out there on the web and say “at this time I consumed another one of those”. I’m not suggesting that everything should have to be described like this, but if there’s a URI to represent something I want to be able to point to it so everyone knows what I’m talking about.
In other words I want daytum.com to be following the Linked Data principles rather than an ajax only interface.
If you have a look at Felton’s own annual reports you will see that they group and aggregate all sorts of information but to achieve something similar (conceptually if not visually) then you will need a lot more from daytum than currently being offered.
The other big gap is the lack of an API to update information. Keeping daytum.com up to date is actually quite hard work and certainly to be able to collect the sort of data Nicholas Felton does to put together his annual reports would be onerous to say the least, but it needn’t be.
If daytum.com provided an API that allowed me to post information from other services that would be a great start, but actually it’s not always necessary, nor even that desirable. The Web already knows quite a lot about us, for example Fire Eagle and Dopplr know where I am/ been, delicious knows what I think is interesting on the web, and how I describe those things, Twitter and this blog what I doing and thinking about; for others Last.fm knows what music they are listening to. Daytum doesn’t need to replicate all of that data, indeed it shouldn’t, it could simply request that data when needed — to visualise it. (it shouldn’t store it because it makes it harder to manage access to it).
The one thing I don’t want, however, is yet another social networking site, I don’t want social features to be part of daytum. I don’t want them because I don’t need them – there are already loads of places integrated into my social graph, whether that be Twitter, Flickr, Facebook or this blog. I really don’t want to have to import and then maintain another social graph. I do however want to be able to squirt the data I’m collecting or aggregating here at daytum into my existing social graph; much as Fire Eagle adds location brokerage to existing services so I want a service that adds personal data to existing social networking sites.