I just came around to getting a Gmail account today. The novel concept appeals to me because I no longer have to use rules to filter email into folders – I simply use filters to produce various views of the same underlying email collection and use the star to notate really important email. Interestingly, it is actually very similar to how I use Outlook 2003 – I use “unread” and Flags to notate important email, and then I watch a few defined search folders for interesting email. Anyway, I will have to give Gmail a couple more weeks before I decide whether it is better or not.
However, one technical detail about Gmail (and Google Maps) really bothers me. For some reason, Gmail and Google Maps insist on instantiating an unused ActiveX control if it detects Internet Explorer as the browser, but it does no such thing on Firefox. Furthermore, the rest of the application does not appear to use ActiveX at all – just the initial detection page.
For example, the following content is the parent frame of Gmail that I get in both Firefox and Internet Explorer (I only altered whitespace characters from the original data to improve spacing and indentation for better legibility, as well as steal colorization from Firefox).
<meta content=“text/html; charset=UTF-8” http-equiv=“content-type”>
<link rel=”alternate” type=“application/atom+xml”
title=“Gmail Atom Feed” href=“https://gmail.google.com/gmail/feed/atom“ />
<a href=”/gmail”>try again</a>.
<a href=”/gmail?ui=html&zy=n”>click here</a>.</p></font>
var fs_time = (new Date()).getTime();
if (top.location.href.indexOf(‘nocheckbrowser’) != -1)
top.location = ‘/gmail?ui=html&zy=j’;
var testcookie = ‘jscookietest=valid’;
document.cookie = testcookie;
if (document.cookie.indexOf(testcookie) == -1)
top.location = ‘/gmail/html/nocookies.html’;
document.cookie = testcookie + ‘;
expires=’ + new Date(0).toGMTString();
var agt = navigator.userAgent.toLowerCase();
if (agt.indexOf(‘msie’)!= -1 && document.all)
“var control = (agt.indexOf(‘msie 5’) != -1) ?
top.location = ‘/gmail/html/noactivex.html’;
“<frameset rows=\”100%,*\” border=0>
<frame name=main src=/gmail/html/loading.html frameborder=0 noresize scrolling=no>
<frame name=js src=/gmail?view=page&name=js frameborder=0 noresize>
Notice that the Gmail parent frame intentionally isolates Internet Explorer’s User Agent string and tries to instantiate various versions of the XMLHTTP control. It does not do anything with the control, but if it FAILS to instantiate the control, it redirects you to the “ActiveX must be enabled on IE” page.
I searched through the content of the child frames as well and found no need for ActiveX. In fact, I configured Internet Explorer to popup on any ActiveX use and browsed every single dialog, and Gmail only asked me that one time on initial load. So, as far as I can tell, Gmail and Google Maps do not use ActiveX anywhere other than the initial “test”, which corroborates with the fact that these applications also run fine on Firefox without ActiveX.
So, the million dollar question in my mind remains. Why do these applications unnecessarily “require” ActiveX on Internet Explore if it is not used?
When I use Internet Explorer, I unilaterally disable ActiveX behavior, so it bothers me that I have to alter my security settings just to get Gmail/Google Maps to work – and then to find out that these applications really do not need ActiveX.
Thus, I am more curious about improvements than anything else, and I hope that Gmail and Google Maps will remove the unnecessary ActiveX probe (or at least explain why it exists).