Every engineering team I’ve been a part of has eventually started a technical paper reading group.
At Yahoo! it was a group of engineers from a variety of teams, and at Digg
it included the entire engineering and operations teams.
A couple months back we started up a reading group at SocialCode,
and since then we’ve discussed Google’s Spanner, Paxos, pretotyping,
Brewer’s CAP theorem and a handful of other papers.
We’ve done a few things differently than I’ve seen in previous attempts, and I think they’ve made the
reading groups meaningfully more valuable.
The first is that we’ve expanded beyond engineers: our research and product teams are also joining the reading sessions.
While this has created some constraints around picking papers with topics and treatments appropriate
to a less engineer-centric audience, it has also allowed us to take on papers that play more to the product
and research teams' experience, including those on network analysis and pretotyping.
Beyond learning from the other teams' experience, this has a beneficial side-effect of strengthening
informal relationships between the teams. Since we’re spread out geographically, this side-effect
is probably just as valuable as the original goal of learning about new techniques and technologies.
Lenses for Reflection
The other difference with previous attempts is that we’re trying to stick to this agenda:
- fifteen minutes on the paper’s contents,
- fifteen minutes of apply the paper’s contents to our work.
The second half tends to be the most interesting part. It gives us a weekly forum for reflection,
each time from a different perspective.
If you believe that learning occurs
during reflection–an idea I appropriated from my father when he was working on
ongoing faculty development at UNCA–then this represents a powerful
The reflection sometimes goes in very unexpected directions, giving this paper-reading
format substance where others have at times felt superficial.
When combined with the short discussion period of thirty minutes,
it’s been lightweight enough for people to attend regularly and valuable enough for them to bother.