Full Idea
Donnellan objects that Russell's theory of definite descriptions overlooks the referential use (Russell writes as if all descriptions are used attributively), and that Strawson assumes they are all used referentially, to draw attention to things.
Gist of Idea
Russell only uses descriptions attributively, and Strawson only referentially
Source
report of Keith Donnellan (Reference and Definite Descriptions [1966]) by William Lycan - Philosophy of Language Ch.1
Book Reference
Lycan,William G.: 'Philosophy of Language' [Routledge 2000], p.27
A Reaction
This seems like a nice little success for analytical philosophy - clarifying a horrible mess by making a simple distinction that leaves everyone happy.