Google's web history application is quite useful. At least for Google searches. To expand web history beyond searches, one must install Google's Toolbar. I don't like browser toolbars and haven't installed Google's, so I only see searches in my history.
About once a month, I use search history to find a page I looked at earlier but am unable to track down again through other means. As the first step towards getting a convenient, browsable history of my browsing, tonight I implemented Hendricks Solutions Web History.
It's just a simple Greasemonkey script that shoots off an XMLHttpRequest with the URI, timestamp and page title of the web page that just loaded. It looks something like this:
// ==UserScript==
// @name Where have I been
// @namespace http://www.ndrix.com
// @description Submit an AJAX request registering every URI Firefox visits
// @include *
// ==/UserScript==
var uri = location.href;
var title = document.title;
var time = Math.round((new Date()).getTime()/1000 );
var get_this = "http://www.example.com/foo.pl"
+ "?time=" + time
+ "&uri=" + encodeURIComponent(uri)
+ "&title=" + encodeURIComponent(title)
;
GM_xmlhttpRequest({
method : "GET",
url : get_this,
onload : function () { return },
});
On the server there's a small Perl script that collects that information and appends it to a log. Now I have a central repository of all the webpages I ever visit. Sometime I'll hack on a pretty front end.