a number of literalist posters who claim that science can't be applied to past events since "you weren't there".
Ned, perhaps part of the problem is nebulous use of the term "science". In most people's minds I would guess that "science" conjures up ideas of an academic class or department across campus from what they would think of as "history". Attempts to explain how history can be a science are thus met with ridiculous comments because many see that two as essentially different, black and white, based on academic preconceptions. Since in simple terms history is the study of the past, the literalist problem you cite is carried over to geology, evolution, etc...
In my mind, "science" is happening any time the scientific method is correctly being applied to a problem. I recently tried to use auto repair as an example of this (and was essentially laughed at by my opponent) - the idea being that though auto repair is generally not considered "science", the scientific method could be used to analyze a novel problem arising with an automobile. Indeed, the scientific method is sometimes used to investigate accident scenes, which brings us around to Crash's example of criminal forensics.
"History" is only a science when the scientific method is applied to its subject - but then it is a valid science, as Mark's quote explains in the OP.
Just as "history-vs-science" is bogging down other threads, so is "misunderstanding-the-nature-of-scientific-method".
Perhaps use of the term "scientific method" or similar should be used in place of "science", which conjures up academic preconceptions in non-academics... however, this would require those same people actually understanding the method behind science.