I was working on a website for a client when we discovered that sometimes the website was not accessible for Internet Explorer. Notice the word sometimes.

A word dreaded by every programmer including myself. In order to solve problems there needs to be a pattern. Something like: “press the red button and the screen goes blank”. Without the ability to reproduce a problem it is really hard to solve the problem or confirm that the actions you took to solve it were indeed the (right) solution and not some fluke.

In this case a user visited the website and got access to it, or the user visited the site and got an arcane error message:

“Internet Explorer cannot open the Internet site http://<Web site>.com. Operation aborted.”

I never saw this message before and after some testing we could confirm it only happened with Microsoft Internet Explorer. There were no problems with any other browser, but somehow Internet Explorer sometimes spat out this utterly useless error message and prevented the user to visit the client’s website.Together with some engineers from the hosting company I sat out to find and solve this issue. So we started to have a look at the web server’s  log files.

At first the engineer from the hosting company suspected it might have something to do with some HTTP 500 error codes, missing images (so you would expect HTTP 404 errors) and some rewrite rules. According to his first thesis this combination might have caused a redirection loop and Internet Explorer could not cope with it. After quickly fixing the missing images (a typo in a path) we could dismiss this thesis and we continued our quest. Luckily a colleague of one of the engineers remembered a similar issue creeping up with a different client of them and pointed us in the right direction. A quick search (note to self don’t try to be smart, first search for the error message on the Net, then analyze logfiles) later gave use quite a lot of links pointing towards Microsoft.

Apparently Microsoft’s engineers think that its a good idea to confront a user of their browser with an arcane error message and prevent the user from accessing the website when the visited website in question does something less smart (maybe even stupid). While any other browser would at least try to let the user access the website and cope with the issue. I wonder where the Interaction Design people were during this decision.

No, I’m not making this up. Microsoft states in the knowledgebase article on this issue: “This behavior is by design.” What are the arguments for doing it this way? Interestingly in the same article they claim:

How do I fix this problem?
The easiest way for you to fix the problem is to upgrade to Internet Explorer 8. This problem no longer occurs in Internet Explorer 8.

So the behavior is by design, yet they advice you to solve it by upgrading your browser. Interesting, to say the least.

Anyway, after reading the article and doing some more debugging (which is a lot more painful then it ever should be using Internet Explorer) I found the culprit to be a piece of JavaScript (swfobject.js in this case) trying to access an element before the parser was ready. A bug and certainly something we needed to fix, but Microsoft did you really have to abort loading the webpage and preven the user from accessing it? Why can all the other browsers cope with this and still allow the user to visit the website? Would it have hurt to show a less obtrusive error message in the status bar?

For those having similar issues please see these urls for more info:
http://support.microsoft.com/kb/927917/
http://blogs.msdn.com/ie/archive/2008/04/23/what-happened-to-operation-aborted.aspx