Not exactly DRM for news

I’m a couple days late with this one. Ars discusses the seemingly ludicrous plan by the Associated Press to improve its news platform online. In which they claim to be able to deliver news online with a ‘tracking beacon’. The whole press release/news story sounded very odd in that nothing they mentioned seemed possible with standard web technologies.

Their poor excuse for an info-graphic has received some hilarious commentary. I think I just like profanity in my parody.

But after some digging it seems to make sense after all; and suspicions confirmed that nothing they originally wrote is actually truthful of what their technology does. The best description was a comment by ‘deet’ on the Ars article (for which the permalink seems broken, so you’ll have to scroll down and find it):

Look at this embarrassing DRM verbiage as a kind of sideshow for the old folks.


AP’s hope seems to be that this new specification for online delivery of AP member content will slow, stop, or at least reveal the activities of the more blatant rippers-off, while giving useful tools to legitimate publishers for monitoring and controlling the use of their content, which is entirely within their prerogative. Obviously, and as with any security system, a sophisticated attacker can circumvent these measures. And the AP knows this. What’s great about the tagging system is, if you’re a legit publisher, the tags had better be there. If the tags are missing, well, be prepared to hear from AP.

(Don’t just read these snippets, the whole comment is longer and interesting.)

Their plan seems to comes down to some HTML metadata that, if present, flags the content as legitimate, and if not (or used incorrectly) yells loud and clear that the text has been misappropriated. The big problems after this information becomes available and widely-used are (a) to get people who are entitled to the content to use the metadata correctly, and (b) to somehow track down non–fair use quotations of the text that aren’t overwhelmed with false positives. (The former being necessary to even have a hope in hell in achieving the latter, assuming that simply looking at the domain name of the hosted content isn’t enough.)

It’s not clear to me that this latter problem is made any easier by the absence of some metadata.