Fastfission wrote:
Perhaps I sound a little territorial, here (of course the historian claims that being a historian is not easy), but I suspect that other people with other backgrounds will agree a bit on this. I'd rather have people write an encyclopedia based on secondary accounts than primary accounts -- the former will prove to be a collection of the current state of the knowledge (an encyclopedia), the latter will prove to be a collection of off-beat, missing-the-point, and thoroughly unaware and uninformed amateurisms.
I suspect you're wrong about this being universal. It may well be true for history, but in mathematics, for example, citing primary sources is perfectly reasonable and even desirable. The difference, of course, is that history, unlike mathematics, requires context and interpretation.
Of course, even in mathematics one should be vary of those primary sources that have not passed peer review and/or public scrutiny. But then, the same applies to secondary sources as well. (People _have_ been known to attempt revisionist interpretations of mathematics. Such kookery is usually much more obvious that historical revisionism, but it does happen. See various long threads in sci.math for examples.)