diff --git a/packages/router/src/router.ts b/packages/router/src/router.ts
index 31971c26e68..1d90f1d55ab 100644
--- a/packages/router/src/router.ts
+++ b/packages/router/src/router.ts
@@ -369,6 +369,21 @@ export class Router {
    * rawUrlTree to be a combination of the urlAfterRedirects (again, this might only apply to part
    * of the initial url) and the rawUrl of the transition (which was the original navigation url in
    * its full form).
+   *
+   * Note that this is _only_ here to support `UrlHandlingStrategy.extract` and
+   * `UrlHandlingStrategy.shouldProcessUrl`. If those didn't exist, we could get by with
+   * `currentUrlTree` alone. If a new Router were to be provided (i.e. one that works with the
+   * browser navigation API), we should think about whether this complexity should be carried over.
+   *
+   * - extract: `rawUrlTree` is needed because `extract` may only return part
+   * of the navigation URL. Thus, `currentUrlTree` may only represent _part_ of the browser URL.
+   * When a navigation gets cancelled and we need to reset the URL or a new navigation occurs, we
+   * need to know the _whole_ browser URL, not just the part handled by UrlHandlingStrategy.
+   * - shouldProcessUrl: When this returns `false`, the router just ignores the navigation but still
+   * updates the `rawUrlTree` with the assumption that the navigation was caused by the location
+   * change listener due to a URL update by the AngularJS router. In this case, we still need to
+   * know what the browser's URL is for future navigations.
+   *
    */
   private rawUrlTree: UrlTree;
   /**