[WikiEN-l] Reliable sources— some of these babies are ugly

[Nathan]

On Mon, May 17, 2010 at 3:27 PM, Ken Arromdee <arromdee at rahul.net> wrote:

> If "riddled with errors" means "has more (frequent) errors than other
> sources", then this makes some sense.
>
> If "riddled with errors" means "has errors that we have recently had our
> attention called to" or "has errors that happen to be about some subject we
> are personally pissed off about", then it's a very bad idea.
>

I agree, and that's why I suggested any decision to "delist" a source
as presumptively reliable be based on an analysis of a selection of
published content. Shmuel wrote that the purpose of identifying
reliable sources is to keep editors from making stuff up -- but we
exclude all sorts of sources that aren't editors making stuff up,
based on a potentially faulty assumption about their editorial review.
So rather than aiming to prohibit hoaxes, rules about RS are an
attempt to weed out chronically unreliable sources. If we find that a
traditionally reliable source of facts has become chronically
unreliable, then it should face the same scrutiny as blogs or personal
websites prior to being cited.

Nathan