PDF Tracking On the Way

(el)Capitan.Nick writes "PDFzone reports that the company Remote Approach has launched a service to track the movement of PDF documents with its tool Map-Bot. The purpose of this service is to allow PDF publishers the ability to measure their audience, as web publishers can already. Though personal information is not gathered from machines, IP addresses are. PDFs can require users to be connected to the Internet in order to read them, and every person you email the PDF to is subject to the service. As PDFzone's opinion article states, while 'the chances of running into a Remote Approach PDF right now -- and in the near future -- are pretty remote ... the potential for the technology to tarnish PDF's image [of security] is staggering.'"

Advertisements!

[Eric(b0mb)Dennis (629047)]

Oh.. soon as they can track views of PDFs, people will start putting ads in them... I guarentee it!

I can see it now.. Google introduces AdWords for PDFs...

Evil, explained

[hummassa (157160)]

Q: How does this tracking mechanism differ from web log analysers?

A: Simple, web log analysers aren't capable of tracking redistributions of the same document. If you copy a web page, say about theories in free-market macroeconomics, and e-mail the copy to a friend, say in China, no one will ever know your friend has read it. But if you copy one of those and it's read by your friend there, then certainly your friend will have a red flag (pun intended) on him.

HTH

Simple...

[Rolan (20257)]

It's simple... Refuse to read PDFs that require the technology. Publishers won't get any data from it, and given a loud enough voice, will find that the tool reduces their distribution. It does them no good if the users won't read their documents because of it.

Re:Simple...

[thedillybar (677116)]

>It's simple... Refuse to read PDFs that require the technology.

You'd have trouble convincing more than about 2% of users to refuse.

>Publishers won't get any data from it

Sure they will. You will be the one getting no data because you're holding out when no one else cares.

It's a wonderful idea, but it simply won't happen without government intervention...and who wants that?

Sure, that works

[John Jorsett (171560)]

It's simple... Refuse to read PDFs that require the technology.

Just like I can shop elsewhere if I don't like being captured on a store's video surveillance camera. Except that they ALL have cameras. If there's no true alternative, you're screwed. Am I going to forego opening that online manual that I desperately need to troubleshoot a problem? I don't think so. A better solution is for some enterprising hackers to find a way to break this technology.

Re:Sure, that works

[John Hasler (414242)]

> Just like I can shop elsewhere if I don't like
> being captured on a store's video surveillance
> camera.

Yes. You can. Contrary to common belief, your choices are not limited to Walmart and Kmart.

But how will you know?

[JoeBuck (7947)]

Disabling Javascript will keep the tracking from working, but if you don't, the transmission is completely invisible to you. It will look like normal HTTP traffic to your firewall.

Re:Simple...

[j1m+5n0w (749199)]

It's simple... Refuse to read PDFs that require the technology.

Better than that, refuse to use pdf viewers that implement this "feature". (Does anyone know which those are? Without knowing, I would assume Adobe acrobat reader probably does and xpdf probably does not. Does anyone have more specific/accurate information?)

As much as it pains me to say this...

[ral315 (741081)]

How is it any different from collecting the I.P. of everyone who visits your website?

Re:As much as it pains me to say this...

[akzeac (862521)]

Websites only collect the IP of the machine that downloaded the page. This technology would distributors to collect the IP of every machine in which the PDF is *viewed*.

On the evil side, getting on the conspiration mood, it would also allow the FBI or the gov to diffund pseudo-dissident bait documents and then check and track anyone who reads it, anywhere he reads it.

Okay....

[Balthisar (649688)]

Okay... Print, Save as PDF on the Mac, or Print, select PDF Writer on Windows, or print to ps and "distill" with gs on anything else, and there goes the tracking. Not right?

Re:Okay....

[Lehk228 (705449)]

ghostscript can read encrypted PDF's, however it does honor the creator settings for disabled features, you will have to go in and recompile it with whatever function checking if it is set to disable features to always return no features disabled.

Disable PDF Javascript

[user9918277462 (834092)]

The remote logging is done through embedded Javascript in the PDF file. Most free viewers such as gpdf, xpdf and kpdf don't support Javascript so you're safe with them.

Adobe Acrobat Reader starting supporting embedded Javascript with version 7.0, although you can disable it in the preferences dialog. Apparently it bugs you every time you start the program to re-enable it, though.

Bottom line: Stick with free software.

Discussed on LWN concerning Adobe Acrobat 7

[nick_urbanik (534101)]

• Article [lwn.net] is subscribers only (worthwhile)
• Article will be readable by guests 1 week after publishing
• Solution in Linux is to disable Javascript in acroread 7

Re:Discussed on LWN concerning Adobe Acrobat 7

[Isthistakenyet? (140224)]

There is a bug (in my opinion) in Acrobat Reader 7 when you disable JavaScript that causes this warning to appear when exiting the program:

This document contains JavaScripts. Do you want to enable JavaScripts from now on? This document may not behave correctly if they're disabled.

This happens even if you do not have a document loaded, since Adobe Reader tries to run some internal JavaScripts when it exits. If JavaScript is disabled, this warning comes up. I've created patches [k-lug.org] that prevent this from happening on both Linux and Windows. They may also prevent the warning from coming up with documents that actually contain JavaScript.

Rather pointless

[hweimer (709734)]

PDFs can require users to be connected to the Internet in order to read them,

No, they can't, PDF is nothing but a data format. Some broken PDF viewers (especially those from Adobe) may do this, but since PDF is an open format, there will always be some other viewers that don't promote spying on their users. Basically, this is the same nonsense as the "no printing" option.

Re:Rather pointless

[NetNifty (796376)]

Depends how it's done though, if it just plain PDF but with javascript as has been suggested so far, then you are correct.

However, if for example the document is encrypted and they key is on a server which the PDF points to (and the server logs all IP addresses connecting to it to retrieve the key) then it will work at least for the first time you open it (unless of course we create another server or even p2p network with the keys on it for ebooks which the PDF viewer visits instead).

They should make another file extension

[saskboy (600063)]

Rather than tarnish the PDF name, they should create the Tracked Document Format or TDF and that way users can distinguish between the two. To make people suspicious of PDF right after versions 5 and 6.0 were found to contain security holes, this will be bad for Adobe.

IP harvesting

[Douglas Simmons (628988)]

I'm going to try to present this in a non-paranoid tinfoil hat mentality: I could see this being a great feature if I were in the PDF sending business for marketing purposes or whatever. Now if I were in the FBI/CIA business, this would be great to use, for example, to proliferate PDFs on Kazaa with filenames/tags suggesting they contain info on how to make bombs or blueprints to the Pentagon so that I could collect IPs of whoever's interested in this type of stuff. You see where I'm going with that.

Also, I definitely do not want to risk exposing my static IP to anyone, especially in a way that involves new technology that may be quite exploitable, just by clicking on a PDF link on google. I'm sorry but c'mon, that's just too much. Nevertheless, assuming the technology is viable, there'll be a demand that will outweigh objection for this new feature and Adobe will do it and make more money.

Re:IP harvesting

[MntlChaos (602380)]

Also, I definitely do not want to risk exposing my static IP to anyone, especially in a way that involves new technology that may be quite exploitable, just by clicking on a PDF link on google

Wait a minute... clicking on ANY link on Google exposes your static IP to the content provider anyway.

Slippery slope argument

[sanityspeech (823537)]

The editor's take on the story makes it seem rather benign. However, the actual story makes it sound more alarming:

Are Your PDFs Spying on You?

Like Adobe Policy Server, Remote Approach can FORCE users to be connected to the Web in order to read the documents. It can track who's e-mailing your PDFs to whom, and what they're reading. Real-time. (Emphasis mine)

FORCE me to go online??? I just hope that technical papers never use this tool.

Denizens of the PDF world, however, take note. We enjoy--and sell--the differences between PDF, e-mail and HTML, and a lot of those differences are in the realm of security...

Remote Approach, however, is the beginning of a movement that could chip away at PDF's sterling rep, one document at a time...

Since the Map-Bot can chase a PDF through e-mail forwarding, it's more powerful data mining than that associated with Web pages, where the vital information gets thrown out when the user's cache is emptied.

One would think they would come up with a better name than Map-BOT!!!

Pretty damning, if I may say so.

Refuse to read PDF's, period.

[AtariAmarok (451306)]

PDF's are great for printing, but not as easy to view on the Internet as regular html files. The Google "viwe as html" tool will help greatly.

A little technical info

[by Anonymous Coward]

Ok, so I downloaded the demo document, and captured the packets.
There's a POST to remoteapproach.com (you could block all traffic going to remoteapproach.com, or just repoint remoteapproach.com to 127.0.0.1 or something in your hosts file.
The POST message looks like:
POST /remoteapproach/logging.asp?type=view&DocID=123456 7890&GroupID=123456789&ChannelID=123456789 HTTP/1.1

The thing that gets me is that the content of the request also contains this:
1 0 obj]/F(/C/Documents and Settings/Administrator/Desktop/MBRemote Approach Manual.pdf)>>>>

As you can see, it contains the full system path to the file that I opened. This seems like a big privacy issue. After all, Acrobat didn't ASK if it could open the URL.

The .PDF files can be opened with Ghostscript, and (obviously) do not send tracking information. Simply re-saving the document as PDF doesn't remove the tracking, but converting it (File--Convert) via pdfwrite APPEARS to remove the tracking.

Some technology.

Re:PDF

[jcr (53032)]

IIRC, it's "Portable Document Format".

-jcr

Re:Thankfully

[GigsVT (208848)]

Not likely, the last change to the PDF license was the ludricrous requirement that all those who implement PDF also implement the "evil bit".. that is the useless tags that forbid you from printing/saving/etc in acrobat (reader).

No one else paid attention to it. Since earlier versions of the spec didn't have the requirement, there's no way they can enforce it. Other than that stupid requirement, the spec has an open and free license.

Besides, only Adobe products implement javascript in PDFs to start with, so Adobe brought this on themselves. No other reader will allow this to happen.