> vulnerabilities that are publicly-known before a patch is available
We already have a term for that, though: "unpatched vulnerabilities". Using the term "zero-day" erroneously communicates to the reader that Google surprised Apple et al with this thing out of the blue, and just handed crackers the tools to start exploiting it without giving Apple time to develop countermeasures.
Indeed, but it was also an 'unpatched vulnerability' for the last 90 days. And, it was even plausibly an 'unpatched vulnerability' before it was discovered!
What's special about today, with an exploit being widely-known but no patch available? That's the new situation people are grasping for a term to describe.
'Zero-day' somewhat fits, at least from the perspective of everyone outside Apple: OSX users and those potentially attacking them. It's a brand-new green-field risk/attack for them!
It's also somewhat problematic, for the reasons you list and others. There may be a better term yet to be discovered, that maintains the special distinction for vulnerabilities that the vendor only learns about simultaneous with exploit-availability. But still a lot of people use and understand the term to apply to the window-of-danger from unpatched "Project Zero" 90-day reveals.
We already have a term for that, though: "unpatched vulnerabilities". Using the term "zero-day" erroneously communicates to the reader that Google surprised Apple et al with this thing out of the blue, and just handed crackers the tools to start exploiting it without giving Apple time to develop countermeasures.