Gerry Carr is head of platform marketing at Canonical and is reporting directly from Ubuntu Developer Summit this week. We're hosting those exclusive reports here as a series of guest blogs. Day 4 summary here...
Last night I joined the bulk of the design team at a Massive Attack concert in Universal City Walk - a truly weird place. The concert was amazing though, especially as they played to a background of binary numbers, which looked to my eye like a very cool launch event for Ubuntu 10.10. Or maybe I've been here too long.
But back to UDS this morning. A theme I have picked up on this week, that I touched on yesterday, is the degree to which we have moved towards building features based on research. I was struck by this at an interesting presentation by Michael Terry of the University of Waterloo that I attended, provocatively entitled 'How to spy on your users'. The title worked in attracting me along.
The goal of the project was to get instrumentation data from users in how they use applications. Basically a user opts in to send data on how they use an application or a system. That data then informs the design of the next revision. For instance, if there is a feature that is never used, is there any point in keeping it?
The discussion looked at the issue of self-selection bias with an opt-in system. Mozilla offered the information that 97 percent of its volunteers for Test Pilot were male for instance, hardly reflective of the actual user base of Firefox.
Collecting instrumentation data is not something we do in Ubuntu as yet, and there will need to be considerable debate in order to design a useful system for doing so. In theory, there is great value. In practice you might get a lot of data telling you that users like to open and close windows and click on icons.
But the theme continued through the plenary. Ivanka Majic, who heads up design, spoke about the dangers of design by enthusiasm. Doing things because you can, can lead to designing Homer's car. A vision needs to be outlined and a philosophy of design articulated. Large participation is great but it needs to be a group going in a similar direction with some shared values. Then you can rock the design.
And on it goes. Maybe I am suffering from self-selection bias myself. In all honesty I don't know if the kernel track, for instance, has been electrified by the outcome of user input. But discussions from accessibility through the software center through multi-touch design all had use cases and research as the primary driver for prioritization. Or if they didn't have that data the next step for discovery of what features to add is to do some form of research from the multiple sources that are available.
Ubuntu and Canonical often get criticized for merely being package aggregators; that is we take the work of upstreams, put a bow on it and call it Ubuntu. But where else is this focus on research into the existing and potential user base for Linux as a desktop OS taking place?
And where else is this research moving so quickly into shipping product?
And of course it is not just Ubuntu that benefits from the application of this learning. The upstreams we ship and the distros that are fellow travelers are positioned to benefit from it, should they so choose.
So users of Ubuntu have a lot to look forward to in this cycle. In the words of Massive Attack, the big wheel keeps on turning...