Here is a quick follow-up of the event I announced in my previous post: the Distributed C++ Meetup 0x02. A quick explanation for those too lazy to click a link or scroll down a bit to read my previous post (not judging you here, I would do the same); the concept of a Distibuted C++ Meetup consists in gathering multiple C++ user groups from around the world in one event using video-conference facilities. This time we had the pleasure to bring together the Berlin, London and Stockholm Meetup groups using King's offices.
Distributed C++ Meetup 0x02:
Once again Phil Nash took on himself to edit a full footage of the event, which you can enjoy right here:
We also have altenative footage of the two first talks made by Harald Achitz:
Being the host of the event makes you easily distracted, and so I was! But thanks to Phil's and Harald's work, I can enjoy a rewatch of the talks in a serene environment.
Harald Achitz gently explained us the space-ship operator that will arrive in C++20 and what to expect from it. Then came the introduction on Type Safe Flags made by Arvid Norberg that I particulary enjoyed : simple and very practical! Richard Spindler related us his experience with Boost.Compute. Finally, Dominic Jones made a talk on Expression Tree Transform which reminded me a lot Reactive Equations by André Bergner that I had the chance to see during Meeting C++ 2017. Although, I always enjoy template meta-programming, it was very alien to anyone at the event who wasn't familiar to expression templates. I always admire anyone willing to be on stage to share her/his knowledge and I appreciate the honor our speakers did to us by doing so that evening.
As the tradition goes, we sent a survey after our Meetup event. The feedback we received was substantially more divided than last time. Although the numbers indicate that a majority of the people are willing to redo such an event ; 47.50% really want a re-run, 35.00% liked it as much as a normal event, 17.50% would have prefer to stick to a normal one; it was clearly less popular than the last event with respectively 82.50%, 15% and 2%. Surprisingly people changed their minds and now would prefer going back to lightning talks (instead of 30min ones), as suggested by few feedback comments. As a comparison you can also read more about the previous event's results on Phil Nash's blog. People also expressed a strong feeling of being less "connected with the remote crowd" this time, which is sadly one of our main goal for a Distributed Meetup. It was especially surprising that this time we had questions crossing the virtual borders.
I have been reflecting on what changed in our event's recipe and what could have be done better. Here are my thoughs as well as observations from other co-organizers:
- The more cities we gather, the less time each location has talks happening physically in front of them. No matter how good is the video-conference setup, anyone would prefer to have it "at home". While as an attendee you will appreciate spending half of the event learning from speakers at another location, two-thirds of the event might be too long. Having more cities also increase the risk of cumulating technical issues. I would suggest to come back to a two cities format.
- We need to have room for a break in our schedule. Even Supertemplateman would not stand two hours of C++ talks on a cramped bench. Having 15min to make our transitions between countries would be pretty sweet too.
- Lightning talks may be somehow frustrating as they barely touch the surface of a topic. But a 35min talk can also be pretty damn long if you do not like the subject, even more so when projected on a screen. We would like to try 15min for a next run of the event. Shorter talks also have a desirable side effect: speakers will tend to write slides that are very expressive. Which bring us to the next issue.
- I noticed that slides were much harder to read especially on the weird 4-TVs setup we have in the Stockholm's office. This was corroborated by the feedback specific to Stockholm. Tiny fonts force you to concentrate to extrapolate information that you cannot obtain with your eyes, it's exhausting and frustrating. We should really advise the speakers to use a minimum font size prior to the event. There is also the possibility to have multiple screens in the Stockholm office, to improve the experience for the people that always sit in our very attractive carousels in the back (every King employee did that mistake once).
- People don't like the mandatory NDA to access the London office. I am not sure how this could be solved, as I am not familiar to this office. Be sure that we will explore that topic later on.
- Finally, as a reminder, never forget to turn the light on! As much as we are used to darkness in Sweden, it was still a challenge for us to see what happened in the tenebrous London office.
I would like to run a third Distributed C++ Meetup with the aforementioned "bug-fixes". I am eager to see if the concept is actually viable or if the first run was simply benefiting from the novelty effect. As we are pioneer in these virtual meetups, it is pretty obvious that we will have to
use our attendees as guinea pigs slowly refine our ideas .
Thanks to all people that helped during this event: Mårten Möller, Vasil Georgiev, Phil Nash, Harald Achitz, Tindy Hellman, Bruno Mikus, our tech and video-crew...