IETester rendering accuracy

August 2nd, 2008

Anyone interested in testing websites for cross-browser compatibility should have a look at IETester. This free tool allows you to test 4 versions of IE (5.5, 6, 6 and 8beta1 (*Update: this has since been upgraded to 8beta2)) all from the same installation of windows.

This is a god-send for people who develop websites in linux (myself included). Before I had a host of virtual machines running WinXP, each one with a different version of IE on it. This may seem a bit bloated, but I was convinced that many of the multiple IE solutions already in the wild seemed alright, but missed some of the subtle rendering problems of the proper versions. It would be awful to think a site was rendering properly in all target browsers, when in fact some subtle things went unnoticed.

I viewed this product with similar skepticism, but the benefits that it would bring me still warranted a look. So I gave it a go, and decided to test the rendering of the Acid2 and Acid3 tests. The advantage of these tests is that they also probe the javascript support for each browser as well as the visual rendering. Accurate rendering of these tests would indicated that IETester is behaving exactly as each browser should.

Results

The results are pleasing:

Acid2

IE Version IETester Reference
IE 5.5 Acid2 Test on IE5.5 using IETester Acid2 Test on IE5.5 reference image
IE 6 Acid2 Test on IE6 using IETester Acid2 Test on IE6 reference image
IE 7 Acid2 Test on IE7 using IETester Acid2 Test on IE7 reference image
IE 8b1 Acid2 Test on IE8 using IETester Acid2 Test on IE8 reference image

Acid3

IE Version IETester Reference
IE 5.5 Acid3 Test on IE5.5 using IETester Acid3 Test on IE5.5 reference image
IE 6 Acid3 Test on IE6 using IETester Acid3 Test on IE6 reference image
IE 7 Acid3 Test on IE7 using IETester Acid3 Test on IE7 reference image
IE 8b1 Acid2 Test on IE8 using IETester Acid3 Test on IE8 reference image

The screenshots show great rendering compared to the reference images. Differences in resolution and viewport size makes some of the renderings look off, but the actual content is rendered properly. The eagle-eyed will have spotted that some of the numbers in the Acid3 tests are different. I believe this is due to the screenshots being taken before the javascript tests were completed. In any case, the rest of the rendering of the Acid3 is spot on for all browsers, so I’m not worried about this little difference.

Conclusions

This is a fantastic tool for cross-browser testing. I have enough confidence in its correctness to now get rid of my entire farm of virtual machines. Not only is it accurate, but it is also very nice to use, with a well thought out user-interface. There are some little bugs in the program itself, but this is to be expected from such a young project. None of them are obtrusive enough to put me off using the program. On the whole I’m very happy to use it—saves me a lot of hassle.

Resources

  • IETester Homepage – Get IETester here
  • Browsershots – A fantastic free service that renders a screenshot your website in a practically every browser in the wild—and then some! (Used for the reference images above.)

Your Thoughts

Comments, questions and suggestions are always welcome. Just leave a comment.

— Martin McNickle

Posted in browsers, web development, testing | 1 Comment »

« Previous Entries