So, look at this graph: http://netmarketshare.com/report.aspx?qprid=1&qpaf=&qpcustom=Firefox&qpcustomb=0&qpsp=200&qpnp=24&qptimeframe=M http://netmarketshare.co

So, look at this graph: http://netmarketshare.com/report.aspx?qprid=1&qpaf=&qpcustom=Firefox&qpcustomb=0&qpsp=200&qpnp=24&qptimeframe=M http://netmarketshare.com/report.aspx?qprid=1&qpaf=&qpcustom... Notice that major bump in August 2016 and after? Firefox 48 was released on the 2nd August 2016. This was the first release to ship the new multiprocess architecture. And this new architecture also already required breaking all extensions. They could not have shipped it this early without knowing that the switch to WebExtensions will cut off the multiprocess-incompatible extensions. And they needed to get multiprocess shipped. Had that graph continued as in the half year before that, we would now have negative user numbers.