Here's Wassom's (the guy I linked to in the prev. item)
excellent summary paragraph on deep-linking:

Deep links by themselves, then, pose little threat of copyright infringement. This point is deceptively simple, but vitally important. Because such links do not appropriate, alter or influence the documents they reference anymore than a reference in a library card catalog misappropriates the content of a book, these links are benign in terms of copyright law. But quite like a library index, they are essential to maintaining the accessibility and utility of the Web itself. A regulatory scheme that attempted to require Web authors to obtain permission before linking to other pages in the name of protecting copyrights would be doomed to backfire disastrously.