Hey Lodewijk,
Apologies for the delay.
This is my second attempt at replying because my first response was just getting bogged down in detail. So I'm going to jump into the crux of the issues.
Given the resources that goes into Wikimania I do personally think that we have an obligation to try to maximise it's impact beyond the four walls of the conference venue. There have long been discussions about what Wikimania should be and the extent with which it could be leveraged as an outreach tool and I think that recording and streaming are both tools to help it reach it's potential. It comes up every single year and we manage to do it better some years than others. It ultimately takes a lot of coordination, some expertise and healthy dollop of time and a strong internet connection (at some point along the pipeline).
This year the WMF, lead by the communications team, decided to experiment with running a fairly lightweight live stream setup of the main conference room and utilising it via social media, banners and a small segment of our donor list. This isn't the first time such a banner has run for a keynote. Something similar was done for one of the 2014 Wikimania Keynotes with the main auditorium streamed throughout in a similarish fashion to this year. We wanted to see what was possible, what could be learned, what should be done differently and whether it was something that we felt would be worth doing more of in the future. This was done with less planning that I would have hoped and as a result there will be much to learn from it and there will be a review of the streaming from this year and the activities around it and I imagine there will be clear recommendations going forward to next year.
In terms of CentralNotice and its usage, there is a quite rightly many a question about whether this or any other usage of it is appropriate. CentralNotice is possibly the most powerful non-profit communication tool on the planet. So powerful that projects like Wiki Loves Monuments can achieve world records with it. Questions that we have, as a movement, never asked ourselves include how it could or should (an important difference) be used, what do we consider an efficient or effective usage, what are our readers interested in, what are our editors interested in, is it possible to give them a choice rather than force them to see everything or nothing? At the moment we lack the data to inform those conversations. I think a good chuck could be solved by simply asking readers and editors what they are interested in what they
Regards Seddon