Friday 14 December 2007

pdf forgeries

It's not every day an email starts like this:

PDF files have essentially become the standard within the business community because of the need for a protected file. However, the software is useless if users need to edit the document in any way.

Oh ho ho (he says, seasonally), does this mean the sender has a way to edit pdfs undetectably? This I must see. The software is called deskUNPDF Professional and it comes from Docudesk. It promises to:

convert PDFs into Word documents, view data in XML-format, or convert the files into HTML for a web presentation

It actually throws the result out in a huge variety of standard formats including images, csv and Sony Reader's lrf.


The danger appears to be that you could 'round trip' a pdf by exporting to word or an image file, fiddle around a bit and then print to pdf using one of the many pdf writers around.


Fortunately for people trying to protect their pdfs, the exercise proved less than satisfactory. Forgeries are evident. So the real value of the software is that you can move a pdf into another format.


Below, I've round-tripped an image page and a text page from the World Wildlife Fund's "Sustainability at the speed of light". It is utterly evident that I've been up to no good. And that, I believe, should be proclaimed as a valuable feature of the software.


Here's the original and the pdf output going via Word:


Compare1


It's shrunk and there's a bit of textual overlap. The Word export has strange column breaks and the different text blocks appear to be in the wrong sequence when editing.


In honour of the Kit Kat tv commercial where the pandas roller-skate when the cameraman's not looking, I thought I'd do a bit of panda substitution using the GIMP.


Compare2


I didn't add the speed streaks (and I cannot repeat the phenomenon) but they do look rather nice. However, even had my panda been better executed, this is clearly no way to forge a pdf.


Bear in mind that a lot of pdfs are protected by copyright and you need to be sure you're not going to land yourself in hot water by republishing. (Hopefully my snippets aren't going to get me into trouble.)

No comments:

Post a Comment