This is a great thread, thank you all. If I squint I can see pretty good consensus around what the biggest issues are, and the list doesn’t seem too horribly long.
As for the timing of 1.0, we have to strike a balance. With a later (and hopefully better) release, julia would be easier to adopt, but we would run lower on resources and possibly miss opportunities in the meantime. In particular, we can’t schedule 1.0 to land on the day we run out of money, so we have much less time to work on it than one might think.
I see the point about tension between semantic versioning (which demands stability) and public perception (which demands cool features). If we’re strict about prioritizing breaking changes, perhaps we can release a 1.0-beta fairly early, giving ourselves more time to work on the debugger and packages before 1.0-final.