When users access a Web site, their computer sends a request to the site's server to begin downloading a page. Each element of a requested page (including graphics, text, and interactive items) is recorded by the site's Web server log file as a "hit." If a page containing two graphics is accessed by a user, those hits will be recorded once for the page itself and once for each of the graphics. Webmasters use hits to measure their servers' workload. Because page designs and visit patterns vary from site to site, the number of hits bears no relationship to the number of pages downloaded, and is therefore a poor guide for traffic measurement.