So I am at SFO waiting to board my plane back to Australia – it’s been a long, long, long eight days.  I arrived in SFO on Saturday morning and spent a bunch of days sight seeing, working, a little writing and then actually putting together the slides and videos for my talk on Thursday.  Over the course of the week I ate loads of Mexican food (and other good foods!), drank some good beer, found a nice dive bar – hi Lillian and other bar flies, visited Borderlands Books, and did touristy stuff like the Castro, Haight, Alcatraz and some shopping. Then came Puppet Camp on the Thursday and Friday.

The Puppet Camp itself was great … actually better than great … excellent.  Not just because of the good technical value (the signal to noise was one of the highest of any conference I’ve been to) but also the good conversations and putting names to faces – hi Brice, Paul x2, Rein, Markus, Beth, Dan, Nigel (“get a dog up ya”), Andrew x2, Paul, Miah, Ben, David, Deepak, Alessandro, Carl, Michael, and probably a dozen others I am forgetting.  Also great to see Luke and Teyo.

The structure of the event was particularly pleasing … structured talks in the morning and unconference sessions in the afternoon – some great topics discussed there too and you’ll see the results of these sessions feed back to the mailing list (some already have been) and reflected in code and feature tickets, for example the Facter “refacter” took a big step forward thanks to Paul and Rein sitting down and nutting out ideas.

The conference was also more than just Puppet – it was a gathering of serious sys admin, ops, development and engineering players who are serious about infrastructure management (and some heavy players in the “infrastructure == code” world) and life cycle.  There were a lot of “war stories” and lessons learnt being shared – I quite frequently overheard “Oh! I never thought of it that way” or “Yeah we have the problem too – this is how we handle it…”.  Even saw some serious “Oh I get it now!” moments – during Brice’s talk on Stored Configuration for example and Paul Nasrat’s discussion on Facter and RSpec testing (“Testing manifests with RSpec? Oh wow…”).

It was also great to see a lot of “dev” and “ops” people talking openly and bridging the (potentially non-existent?) divide between their roles in the life cycle – it is through these sorts of interactions between sometimes divergent views that progress is made and solutions get generated.  The differing perspectives on the nature of the problems themselves also sparked some really spirited and thoughtful discussion – I expect to see some interesting blog posts come out of the conference.

Lastly, we even managed to push some code … stored Configuration using Oracle XE will be in the next major release – “Rowlf”.

I am greatly looking forward to the next one … quick lessons learnt for me would be: somewhere closer to Downtown, skip dinner on the first night, keep the 50/50 structured/unconference|OpenSpaces model, and that SFO is a good place for it.

Thanks for the team at Reductive for organising it and well done!