picture of me wearing old-timey test pilot goggles

After months of work, I’m proud to say that we’ve finally released the second Test Pilot study. It’s called “A Week in the Life of a Browser” (thanks to Jinghua for the catchy name). It is a “basic panel”, designed to run periodically and collect a wide range of basic data about browser performance: session lifetimes, crashes, bookmarks, downloads, searches, and so on. Here’s a detailed list of the information to be collected and some of the questions we hope to answer using this data.

“A Week in the Life” is broad, but shallow. It is meant to be complemented by other, more focused experiments that will probe more deeply into particular areas of interest.

To write “A Week in the Life”, it proved necessary to first release a new version of the Test Pilot extension, version 0.4. If you still have an older version, you’ll have to upgrade to run this study; otherwise, all you’ll see is a blank rectangle. (The upgrade should be offered to you automatically; you can also download the latest version here. )

Once you’ve got TP0.4 and the study is running, you’ll see something like this on your Test Pilot status page:

Graph of bookmarks and browser usage data over time

The top line on the graph shows how my number of bookmarks has changed over the past few days. The bottom bar shows when Firefox was being actively used (orange), when it was running but idle (yellow), and when it was not running (white). I would love to hear your feedback on this graph. Is it readable? Is it confusing? Is it useful? How can it be made better?

Bookmarks and running time/idle time are just two of the types of data collected in the “Week in the Life” study. Other data collected, but not yet graphed, includes:

  • How often do I do searches?
  • How many extensions do I have enabled or disabled, and how does that trend over time?
  • When and how often has Firefox crashed?

I hope to add these datasets to the graph soon.

There are also other types of data collection that we’d like to add to the study: for instance, how much memory is Firefox using? Being able to see at a glance how memory consumption changes along with the length of time the current Firefox session has been running could be very informative — not only for the researchers, but for the users as well.