It is just 2 dates in a year when they try to mark a point.
One is occasion to show new stuff in Blender to industry by releasing a demo.
The other one is a celebration of community allowing development and a way to thank it by releasing a new blender.
They just try to make release of a stable new blender match with these two events.
But it will not be correct to qualify theses dates as crucial events having an impact on schedule of ongoing development.
The announcement for Siggraph was just a beta. They could not achieve this goal. They released an alpha.
It is also common that a release is not ready for blender conference and an RC or a testbuild is done at that moment.
Everybody agrees that is better to delay an official release of Blender to make it stable instead of releasing an unstable release at the event.
The amount of work to put in 2.80 was probably underestimated by euphoria at the time of codequest.
2.8 is such a big change and blender is doing so much things.
Choosing a plausible date would probably require a week of inspection of all tasks progress.
Instead of that, each dev announced an estimation based on his tasks at dev meeting.
It works for next 2 or 3 months. But for a longer period, it is difficult to have an idea of how many bugs, crashes, reviews ; they will have to solve.
2.5 development took years. So, I did not took seriously 2.80 announcement for conference.
I thought that if they could achieve this goal, 2.80 would not be satisfying.
I prefer that they adopt a planning that let them breathe, now, rather than disappointing everybody in September.
It is not true.
Because it is revolutionary, you have no idea of the amount of work it implies.
Fixing bugs, crashes and polishing features… what does it mean ?
It means updating libraries, modifying your plans, sometimes redesigning features or abandoning them for new ones to create.
Actually, there are more than two dozens of confirmed reports about crashes. (They always try to make an official release with no crash at all.)
Currently, bugtracker is closed for bugreports. You have no idea what level it will reach when it will be open.
And on this forum, discussions about 2.8 are not just always about polishing features.
Sometimes, shared views are about completing them (missing basic steps or cool additions), removing them and adopting another paradigms, about blockers, limitations, comparisons with other software behaviours, …
2.8 is an alpha and it is a status corresponding to reality.
In 2 months, it will be a beta. Number of crashes will probably be a lot inferior but bugtracker will be open to bugreports.
There probably will be several hundreds of bugs to fix at that moment ; and still some things from 2.79, not yet, supported by 2.8.
Because 2.8 has to be innovative, there is no reason to keep things missing from 2.79, the way they were in 2.79. Which means that there is room for new stuff in this field of things missing from 2.79.