Ray Saintonge wrote:
Er, a primary source would be the /first/ place a new fact is published (such as a journal). Sorry, but we are not such a place.
That's not the usual meaning of primary source. It generally refers to documents directly related to an event, original manuscripts, archival material, etc. Any written book or article based on these materials by a person who was not directly involved is a secondary source.
This seems to vary by field. In a field such as classics, journals are indeed secondary sources, containing discussion of primary sources (which would include things like Aristophanes's plays). In scientific fields, journals are generally primary sources: a documentation of novel research written by those who performed the research. In this context, a textbook or encyclopedia article on neural networks is a secondary source, but the classic journal article by Kohonen documenting self-organizing maps is a primary source. And I think that's the sort of stuff we're trying to avoid.
-Mark