Microsoft has started a campaign to drive down the market share of IE6. Getting rid of IE6 is a righteous goal. Newer browsers offer better features for user and are less costly for Web authors to develop content and apps for.

Microsoft’s proposed solution isn’t righteous, though. The proposed solution is upgrading to IE8. IE9 runs only on Vista and Windows 7. Those systems haven’t had IE6 available for them in the first place.

There are two main reasons to run IE6, and upgrading to IE8 is a bad remedy to both reasons. The first reason is conservatism and apathy: Just never upgrading anything until you are forced to. The second reason is that the user relies on an intranet app (or in the case of China, on a government or bank Web app) that doesn’t work in anything except IE6.

For those who just haven’t gotten around to upgrading but who aren’t dependent on an IE6-only intranet app, IE8 is a bad target for upgrade, because with the imminent release of IE9 (but not for the versions of Windows that came with IE6), IE8 is about to become yet another legacy version of IE9. Meanwhile, other major browsers—Firefox, Chrome, Opera and Safari—all still have their latest versions running on Windows XP, and at least Firefox, Chrome and Opera are still even getting hardware acceleration improvements on XP. Firefox and Opera even run on Windows 2000.

For a user who isn’t dependent on an IE6-specific intranet app, upgrading to Firefox, Chrome or Opera would be a much better choice than upgrading to IE8 or having to buy new hardware to upgrade to Windows 7 and IE9.

But just saying “Woo! Let’s tell people to upgrade to Firefox/Chrome/Opera!” doesn’t address the problem of having intranet apps that only work with IE6. By definition, such apps don’t work in non-IE browsers but also don’t even work in IE7.

Now, why don’t they work in IE7? Microsoft is famous for bending over backwards in order to preserve compatibility and IE8 and IE9 have numerous modes for emulating old versions of IE. The thing is that IE6 standards mode is not one of the modes available in IE8 and IE9! When Microsoft developed IE6, they kept the behaviors of IE 5.5 available as the quirks mode. However, when IE7 came around, IE 5.5 was kept as the quirks mode but the standards mode of IE6 was replaced by the standards mode of IE7. This meant that any intranet apps that triggered the standards mode in IE6 were in danger of breaking if they depended on behaviors that changed in IE7. That’s why some corporate environments in particular put a block on upgrading IE.

The conclusion Microsoft made from the outcome was that new IE releases should keep snapshots of the engine as seen in the past releases. Hence, IE8 keeps around the quirks mode (IE 5.5) and IE7’s standards mode while adding two modes of its own. IE9 then keeps all the modes that were in IE8 and adds some more.

Now, this might look almost like I was advocating the inclusion of IE6 standards mode in a newer IE. I am not. I think the proliferation of modes in a Web browser is bad for the Web, because if Web sites can choose any one of a large palette of modes, it becomes prohibitively expensive for competing browsers to reverse engineer and clone all the modes.

The original benefit in the competition against Netscape was that when the intranet of a corporation required IE, the users would use IE for all their browsing of the public Web, too. Putting an IE6 mode in IE10 on Windows 7 would, therefore, again expand the intranet-side lock-in to the public Web, which would be bad for the competitive dynamics of the Web browser market.

However, in some cases on the side of the dark matter of intranets, the Browser Wars 1.0–era lock-in to IE6 has become too much even for Microsoft. Those who are locked in won’t even upgrade to a new version of Windows because of it! If Microsoft gets IE6 users to upgrade to IE8 on XP, those users can later migrate more easily to Windows 7 and IE9 or later that contains the modes that IE8 contains.

Short of bankrolling various Windows customers to get their intranet apps freshened up, here’s what Microsoft could do: They could extract IE6 from Windows XP into a regular Win32 app, restrict it to accessing only hosts on an administrator-maintained list of hosts and make it available for free to anyone. This would enable the Win32 backward compatibility functionality on Windows 7 (or Wine which might be a reason they aren’t doing this!) take care of keeping the legacy code running. Links pointing to outside the administrator-maintained list of hosts would open in the user’s default browser, so the intranet lock-in wouldn’t propagate to Web browser choice. Having a short list of legacy intranet app hosts would also mean the legacy engine of IE6 wouldn’t be exposed to random attack pages on the wild Web. This way Intranet Explorer wouldn’t need to be patched against all the threats that browsers exposed to the Web need to be patched against.

Update 2011-06-24: There is an independent vendor that provides glue code that allows Windows 7 users run the engine of IE6 for intranet sites.