Ever since browsers started multiplying, Web designers have wanted to take advantage of new capabilities without breaking the page for users of older browsers. To use image rollover effects safely, the HTML code has to be conditionalized for different browser capabilities. So developers have been coming up with ever more sophisticated browser-detection scripts. For example:

 var browserlevel = -1;function getlevel() {if (browserlevel != -1) return browserlevel;    if(navigator.appName.indexOf("Netscape") != -1 || navigator.appName.indexOf("Mozilla") != -1){        if (navigator.userAgent.indexOf("Mozilla/2.0") != -1)             browserlevel = 2;        else            if (navigator.userAgent.indexOf("Mozilla/3.0") != -1)                 browserlevel = 3;        else                if (navigator.userAgent.indexOf("Mozilla/4.") != -1)                    browserlevel = 4;                else                    if (navigator.userAgent.indexOf("Mozilla/5.") != -1)                        browserlevel = 5;                    else                        browserlevel = 1;        }  else {                  if (navigator.appName.indexOf("Internet Explorer") != -1){                     if (navigator.userAgent.indexOf("Mozilla/4.") != -1)                      		browserlevel = 4;                          else	                            browserlevel = 1;                     } else                      	browserlevel = 1;               }return browserlevel;}

The problems with this script are that it breaks all too easily with a fork in the road as happened with the release of Mozilla source code. A simpler and more robust approach is to use the Document Object Model (DOM) to test for a certain feature. For instance, a one line JavaScript can test if a browser can handle image rollovers:

 if (document.images) {	// code here for browser supporting rollover image effects}