[1] Mostly. I could write a whole "falsehoods programmers believe about programming" about the cases where that model doesn't apply.
[2] Again, mostly. Updates take time, old versions persist, etc.
[3] A few, I'm sure, but that doesn't affect the point.
Also, biology and biomedicine is just as bad, so don't act like like the super-smart physical scientists are somehow superior. Rigor in social sciences is difficult because there are no solid axioms as there are in physics or math to guide hypotheses.
Folks with your type of opinions act like there are such axioms, but that is just simply wrong. This kind of thinking is exactly what I was talking about in my above comment.
No, I agree with you that there are no such axioms which is why the value of social "science" is near zero.