A colleague of mine sent me this link to the Australian Financial Review website. He said they use some Digital Rights Management (DRM) trickery to prevent anybody from copying the story. I thought at first they were just using some script to try and prevent copying but then again this is a huge site, if they wanted to prevent copying they’d do a pretty good job.
I arrived at the site, and selected the text in the article to see if it was true and you could not copy it. Sure enough, I tried to copy and the text that was selected appeared to be gibberish. Selecting and copying a paragraph results in this:
"I c n e p o l m. o k- i e a a c r a l h s o e a a e a p o r a e y n b g v n p o l t e e e h o o i s t e l y o s u t e n s a k n h m o e b e o o t o t a ," W f s y .
I was thinking they must use some javascript to encode the contents of the document when selected but disabling javascript still got the same result when you copy and paste. Pretty clever, and how did they do it? Looking at the source you could see that every second letter in the article was replaced with which is a HTML encoded space. Looking further there was some script that was hiding some elements that appeared to be the content.
It looks like the way they have done it is with CSS, and replacing letters with spaces for every second letter, one copy of content has spaces on even numbers, the other on odd. Line the two up with CSS and viola, readable content! If you triple click a paragraph in Internet Explorer as shown below, the trick is revealed, some letters on one layer select in white, the others in black.
It’s a kind of one way hash, because once encoded, the decoder can not know what is a real space and what is not so it would be tricky to write the decoder.
Some demo code to show how it's done:
<div style="float:left;width:200px;font-family:courier;">C n e t i h r </div>
<div style="float:left;width:200px;margin-left:-200px;font-family:courier;"> o t n s e e</div>
Here's how it looks:
C n e t i h r
o t n s e e
Go on, try to copy it... The only issue is we can't do it with any font other than those where every character is the same width, known as Monospace fonts.
Here's how it looks in the normal verdana font used on this site:
C n e t i h r
o t n s e e
Oops! Either way it's a good way of the AFR preventing people from plagiarising their content with a simple copy/paste operation, and while we can clearly see exactly how it is done, it would be more than trivial to actually decode once it was constructed.
There are some suggestions on this forum thread which appeared after this DRM was implmented. The most workable suggestion there is print screen, it proves DRM can never truly work and in this case it stifles accesibility. The site interestingly enough is unreadable on a mobile phone, even a very recent Nokia N95 model with a powerful browser, because the phone does not have the monotype fonts all the content looks like the verdana font sample above.
