Good points. I would just add two things. First, as you say, the percentage you need to reach herd immunity depends on a lot of factors. Broadly speaking, the vaccination coverage you need is 1 - 1/R0, where R0 is the basic reproductive number of the disease (http://en.wikipedia.org/wiki/Basic_reproduction_number). For diseases like measles, where the estimated R0 is above 10, you need at least 90% vaccinated, perhaps even more; for other diseases it may be lower.
Second, even if you reach herd immunity, it is a concept that completely relies on a random vaccination distribution. Say your R0 is 10, and you have attained a vaccination coverage of 90%, you may still get lots of outbreaks, because the 10% that are unvaccinated may not be randomly distributed. Indeed, it's becoming quite clear that in almost all cases unvaccinated people are clustered. This is why you get these sporadic outbreaks in under-vaccinated communities.
We (some of my colleagues at the Center for Infectious Disease Dynamics and I) are doing a MOOC this fall on exactly these topics: https://www.coursera.org/course/epidemics