Google now officially supports the use of the rel=canonical attribute within HTTP Headers.
In 2009, Google, Yahoo and Bing introduced the canonical tag as a way to allow webmasters to do 301 redirects without physically redirecting humans. A few months later, Ian Macfarlane tweeted asking for a method to do this over the X-robots protocol.
That is what this does - in a sense.
Since you cannot stick a rel=canonical attribute within a PDF, DOC or other file formats, outside of an HTML page - you can use the rel=canonical within the HTTP header to communicate the redirect.
Webmasters are cautiously concerned about the tag - but I honestly think it will be used well for the most part.