On 3/1/07, Rob Smith wrote:
On 3/1/07, Ray Saintonge wrote:
>Steve Bennett wrote:
>
>
>>"Attributability" is much cleaner.
>>
>>
>It looks like some kind of semantic game. Capable of being attributed?
>It suggests that we don't need to make an attribution, only assert that
>it is possible. Actually attributed statements (or attributions) are
>then verifiable.
>
>
The concept is simple and clean. All material published by Wikipedia
must be attributable -- that is, it must be possible to attribute the
material to a reliable source, which tells us it's not a Wikipedian's
original research. But not all material must actually be attributed.
It needs a source only if it's challenged, or if it's the kind of
thing that's likely to be challenged (including contentious material
in BLPs where sourcing is particularly important), and if it's a
quotation.
This way of formulating policy helps to clarify that not every single
sentence in Wikipedia ("the sky is blue") needs a source. Editors have
to use their common sense to ask themselves "is this something that's
likely to be challenged?" and if the answer's yes, they should add a
source.
Forget the issue of "verifying" whether material is true. That's
entirely unconnected to checking whether it has already been
published, which is the only thing the policy's concerned about.
I don't think that we disagree on the principle objective. As I see it
"verify" and "attribute" both accomplish this.
Ec