Glossary

This is a glossary of terms used in the website.

AdTech

Term used to describe advertising technologies and the companies and vendors working in AdTech. Many AdTech vendors utilize cross-site tracking to exploit user data in browser storage.


Analytics

Industry, vendors, tools, platforms, and methodologies grouped together by the pursuit of empirical data for measuring things like web performance, user behavior and advertising efforts.

Often focuses on data collection in a first-party context, but especially in AdTech, analytics tools are used for collecting and analyzing data in third-party context as well.


Chromium

An open-source web browser project, used as the foundation of a variety of browsers (e.g. Brave and Edge (Beta)).

Much of the technology in Chromium is also used in the Chrome browser, but the two projects diverge when it comes to the full tech stack.


Conversion

An event or metadata of a user, collected by analytics tools and AdTech platforms to calculate the efficiency of campaigns and of content optimization and personalization efforts.

Each organization can have their own interpretation of what a conversion is, but typically it is something that has a measurable impact on the business bottom line.


A key-value pair (e.g. userStatus=logged-in) stored in the user's computer. Cookies are keyed to the domain on which they were set. Accessing cookies in same-site requests occurs in first-party context, and such cookies are termed first-party cookies. Cross-origin access happens in third-party context and these cookies are called third-party cookies.


Data management platforms (DMP) and demand-side platforms (DSP) can exchange user information in their own data stores, mapping each platform's own list of users to other lists by utilizing shared keys in third-party storage.

This way a DMP can build a comprehensive graph of a user's cross-site navigation on the web without actually having their own data collection architecture present on all the websites.


Cross-origin

“Origin” is the protocol (e.g. https), hostname (e.g. www.domain.com) and port (e.g. 4097) of the URL. If the origin of the site where the request was sent from differs from the origin of the site the request is sent to, the request is considered to be cross-origin.

For example:

  • A request from https://www.domain.com/page.html to https://www.domain.com/images/image.jpg is same-origin, as both the source and target share the origin.
  • A request from https://www.domain.com/page.html to https://images.domain.com/images/image.jpg is cross-origin, as the source and target do not share the origin.

Cross-site

Cross-site is a type of cross-origin communication, but it's more rigid as it requires that the source and target of the request do not share the effective top-level domain plus one part (eTLD +1).

For example:

  • A request from https://www.domain.com/page.html to https://images.domain.com/images/image.jpg is cross-origin and same-site, as the eTLD+1 (domain.com) is shared by the source and target.
  • A request from https://www.domain.com/page.html to https://imagesource.domaincdn.com/images/image.jpg is cross-origin and cross-site, as the eTLD+1 (domain.com and domaincdn.com) is different between the source and target.

Cross-site tracking

Cross-site tracking refers to a tracking domain harvesting data from user's navigation and actions on other, unrelated sites. This is typically done by storing an identifier in a cookie on the tracking domain, and communicating with the tracking domain in a third-party context.

Cross-site tracking happens covertly, and the user typically has no knowledge of all the data that has been collected from them while browsing the web.


Domain hierarchy

Domain (or DNS) hierarchy refers to the domain names in a given domain name string (fully qualified domain name). With the document.cookie JavaScript API, the browser can read and write cookies on the current domain and all the domains higher in the hierarchy, all the way to eTLD+1.

Thus a user browsing a site on sub.blog.domain.com can read and write cookies on sub.blog.domain.com, blog.domain.com and domain.com.


Enhanced Tracking Protection (ETP)

The name of the tracking protection mechanism in the Firefox web browser.

ETP uses the Disconnect.me lists to determine whether a given domain is a known tracker or not.


eTLD+1

Effective top-level domain plus one part. eTLD would comprise the top-level domain (e.g. .com and .uk) and sometimes a second-level of hierarchy (e.g. .co in .co.uk or .com in .com.au). Thus eTLD is the same thing as the Public Suffix.

The “one part” is then the next level in the domain hierarchy, i.e. the domain name the site would have acquired to map to their servers.

The eTLD+1 term is used in particular by Safari's Intelligent Tracking Prevention documentation to refer to the highest domain name the browser has read/write storage access to.


First-party

In the context of storage access, first-party references the domain the user is currently on. Browser storage is typically completely restricted to the first-party.

Cookie access is a bit different, as the browser can access cookies in a first-party context on the domain they are on and on any domain higher up in the domain hierarchy (all the way to eTLD+1).


Intelligent Tracking Prevention (ITP)

The name of the Safari browser's tracking protection mechanism and related tools.

Due to its wide-spread use (especially on mobile devices), its algorithmic evaluation of tracking domains, and its impact on first-party storage, ITP has been a big disruption on AdTech and analytics industries.


The act of adding URL query string parameters (e.g. https://www.domain.com?id=12345 or hash fragments (e.g. https://www.domain.com/#id=12345) to outbound links from a website with the purpose of passing some key-value pairs from first-party storage to another site without having to worry about third-party storage access restrictions.

Since the values are passed in the URL, the target site can access them by simply taking them from the URL string.


PageGraph

PageGraph is a tool built for the Brave browser, which analyzes the way in which a web document operates (and is operated on) within the browser.

PageGraph builds a representation of things like node changes, network requests, and script execution, so that this research could be used to, for example, optimize the privacy-preserving feature restrictions of the browser that produced the graph.


Quantum

The browser engine used by the latest versions of the Firefox browser. It's an improved / redesigned iteration of Firefox's original Gecko engine.


Referrer

When the web browser issues a request to a web server, that request often includes the referer [sic] header to indicate the URL the request originated from.

If the target resource is a web page, i.e. the request originated from a link click, when the web page is rendered the referrer string is written into the document.referrer property.

Due to its capability to carry information that could be utilized in cross-site tracking contexts, browsers are actively working towards making the referrer string less informative by a process called referrer downgrade or by completely stripping and/or spoofing it.


Referrer downgrade

Downgrading the referrer means stripping out parts of it that could be utilized for cross-site tracking or carrying sensitive information.

Browsers are working on defaulting to strict-origin-when-cross-origin, which would mean that for request outside the current domain namespace the referrer would be stripped of its path, query, and fragment parts. Thus https://www.sourcedomain.com/some-page-with-link/?id=12345 would become https://www.sourcedomain.com in the referer header.

The Brave browser strips the referrer in all navigational cross-origin requests (e.g. clicking on a link to move to another page). For other cross-origin requests, the referrer header is spoofed to contain the origin being requested rather than the origin being referred from.


Script-writable storage

Refers to browser storage that can be accessed with JavaScript. Typically this would be browser cookies, localStorage, sessionStorage, and IndexedDB.


Third-party

In the context of storage access, third-party means that the browser is trying to access storage that is not on the current domain.

For example, while on web.domain.com if the browser makes a request for an image on image.imagestore.com, the request would have access to cookies written on image.imagestore.com only if the browser did not block third-party cookies. Similarly, if the response from image.imagestore.com tries to write a cookie with a Set-Cookie header, this would only work if the browser allowed third-party cookies.

If the web browser loaded another domain's content in an <iframe> element, then any storage access (using e.g. document.cookie or localStorage) within that <iframe> would happen in a third-party context.


WebKit

A browser engine developed by Apple and used as the engine for the Safari browser and all the iOS browsers.