A user should be able to verify that the contents they see when navigating directly to a story match the contents of a particular version of an article. They should also be able to detect when they are not equal. This could be achieved by describing a methodology such as performing a particular hash on a particular content, such as as a SHA-1 on a properly normalized text of the article. This probably couldn't be the HTML of the URL, since things like ads, "current time", or suggested stories could change.
A user should be able to verify that the contents they see when navigating directly to a story match the contents of a particular version of an article. They should also be able to detect when they are not equal. This could be achieved by describing a methodology such as performing a particular hash on a particular content, such as as a SHA-1 on a properly normalized text of the article. This probably couldn't be the HTML of the URL, since things like ads, "current time", or suggested stories could change.