|
The Cavitand Site
Free Resources for a Free Internet |
Home
Documentation
|
Projects | Tools | Docs | About |
P3P is a simple, automated way for users to gain more control over the use of personal information on Web sites they visit.
It has been developed by the World Wide Web Consortium (W3C), and the P3P 1.0 Specification has been realeased just a few days ago (16/04/02), although previous version of the document have been around for a long time.
You are to visit www.p3pexample.org. You type the address in your P3P capable browser and wait for the page to be showed on your screen. Meanwhile, your browser is trying, BEFORE accessing the site pages, to locate and analyse the P3P policy of www.p3pexample.com. Somehow (more below), it manages to get the P3P XML files. They match your preferences, so the page is retrieved and you get the page you wanted. To you, it looks like nothing special has happened. You continue your visit, and then decide to purchase something. The site has a different policy for the /purchases directory, and it doesn't match your preferences. The browser stops retrieving that page and ask for confirmation.
Simple. You have to do nothing. At least, that is the main idea of this project. Sites publish their privacy policies in a machine-readable way, and your browser should retrieve them and decide, based in your preferences, if you want to visit that site, or not. Of course, previously you have to define some items that will configure your online privacy concerns and will allow your browser to assume decisions on your behalf.
There are at least 3 files a site have to create to store P3P policies:
The reference file is something like a map: it defines which policy file (or part of a policy file) applies to which part of your site. For example, you can have a different privacy policies for your /public and your /private directories. Of course, you can apply the same policy to all part of your site.
The XML policy file is the machine-readable specification of your site policies. It comprises what you collect aboout a user, the purposes of that collections...
The human-readable policy file is just the speficication of your privacy policies.
P3P Important note: although P3P is a way to communicate the privacy options of a site, it does NOT ensure the site is compliant with their stated policy.
The first thing a browser (or user agent) has to do when visiting a new site is to locate the reference file. There are 3 ways for a site to let a browser know where the reference file is:
Once the browser has obtained the reference file, it should study it to determine where your request belong (which part of the site), and which policy file (or part of a file) applies to that part according to the reference file.
The next step is to retrieve the relevant policy file. Once obtained (if possible), the user agent will parse it and match it against your preferences. Depending on the correspondence between the information the site gathers and your privacy options, the browser may determine the site is suitable or not. The actions in case they don't match are not specified, and each browser can inform you in a different way.
There are already some very useful tools related to P3P. For example:
Here are included links for the three P3P files in this site. We hope they will clarify some ideas about P3P: