History of Science vs. Chinese Whispers

[Update: Maybe I should add that I do not mean any purposeful making of myths and legends, just a diametrical difference in the approaches of historians and researchers to old sources. Historians, IMHO, assume that a context existed in which the source made entire sense. That seems analogous to economists' rational choice theory (they assume an economic rationality behind even the apparently most irrational choices). With that approach historians should be particularly interested in understanding the passages in an old source, which at first sight seem nonsensical. Researchers tend to skip those passages and instead cite the other ones which they can still comprehend from within their current context/frame of interpretation. The conflict between historians bashing bad retrospectives and researchers proposing them seems programmed. And I'm not talking about urban legends of the kind of an apple falling on Newton's head, but of stuff one can find int he introductions of peer reviewed papers, for example, citing the “Weismann-Muller-Fisher theory” as though they were of one piece.]
 
The other day I wondered why history of science was constantly at cross purpose with the myths and legends emerging from science. It occurred to me that all students at one point in his research project (some on embarking others on writing up the results) turn to the literature of their field with the question whether and how it speaks to them (or on their issue).
     This is fine research but not fine history, because it interprets the source literature within the current context of the researcher's question or problem. For an analogy think of a game of Chinese Whispers with the complication that most of the earlier whisperers in the queue are dead and cannot be personally asked what they meant.
     The role of historians of science, as I understand it, is the reverse. They try to undo the myths and legends emerging from the Chinese Whispers of research citation (for examples look at The Renaissance Mathematicus or PACHSmörgasbord).
     Unfortunately, the conflict is internal to the system, because that what is good science is often bad history. As a research student you are often expected to trawl as deeply as possible for research relevant to your problem. Unless you are also a good historian, however, you will probably miss the fine differences in the context around old sources that give it a different meaning from your interpretations.