On Wed, Jan 7, 2009 at 12:58 AM, Carl Beckhorn cbeckhorn@fastmail.fm wrote:
On Tue, Jan 06, 2009 at 07:44:58PM -0500, WJhonson@aol.com wrote:
It is *already* covered in a text. In fact, I note, just on Google Books, at least six print secondary sources which *mention* it, and a few go into details.
A book which only mentions a theorem but doesn't go into depth is useless as a source. I would always cite the original paper in preference.
Why not both? Wikipedia requires editorial judgment for some things, but selection of primary sources is one of the more tricky ones, and a secondary source showing that you are not cherry-picking the primary sources is a good safeguard.
What I was suggesting is that an article with no secondary mentions (of any kind, whatsoever) is probably a good AfD candidate.
Every topic I am intrested in having an article for will some sort of oblique secondary mentions - but I don't consider those to be sources for the article, and would not include them when I add material.
Consider those oblique secondary sources to be "notability sources" to "allow" the use of the primary sources.
Carcharoth