I develop a web application and we test it on several browsers, including Firefox and Internet Explorer. During our development we tested the application on Firefox 3.5.2 and our testing team tested the same on 3.5.6 and found the UI looks good. Once in production, though, we've started receiving client complaints that it doesn't look good in 3.5.8.

How do we handle these minor version issues? How can we check the browser compatibility during development and during testing?

Is there any tool to test the application on 3.5.X, X being all the minor versions?

  • This is exactly why I stopped making web applications. The only advice I could give you is try to isolate the browser-specific stuff in such a way that you never have to edit your "good code". This can be quite hard, as each browser behaves different at so many different points...
    – deltreme
    Sep 27, 2010 at 11:57

4 Answers 4


I would say stick with the age-old traditions in web development.

Stick with the standards. Add small adjustments for the specific browsers your users are using if necessary.

You can't really do any better than that. There is no tool that will ensure your site works on every single browser and version. Stick with the standards. Add small adjustments.


Unit Test Your Webapps:

There are several tools that let you check major and some minor versions of websites by comparing screenshots and unit testing frameworks that let you automatically drive the browser and both content AND functionality are correct. If time is an issue, I would suggest option #1. If you have time to write unit tests, which is the better option, then go with option #2.

  1. Try http://browsershots.org/. There are others as well that do the same thing. Without much work at all you can do spot checking of hundreds of browsers and versions.

  2. If you want to get more detailed, I would suggest writing web UI unit tests that give you more control over your testing. Great libraries include Selenium (C#, Java, Ruby and many more), or WatiN.

  • 1
    I think you're misunderstanding the concept of 'Unit testing'. Automated testing != Unit Testing. Sep 27, 2010 at 13:11
  • Writing Selenium code in a JUnit unit test isn't unit testing? (I agree browsershots.org isn't unit testing.) Maybe I tried to answer too much in my question and it looks bad from merging the two options.
    – Ryan Hayes
    Sep 27, 2010 at 13:29
  • 1
    Writing Selenium code in a JUnit unit test is not unit testing. It's integration/system testing being run by a unit testing platform. Unit testing means you test a single unit of code in isolation, for example a single method on a single object. Browser testing is NOT unit testing. Sep 27, 2010 at 18:42

This is one solitary confined case where I would say IE has the edge over other browsers.

Version released. BAM! Nothing till the next version. It has issues? Work around them. The key here is the issues are consistent and repeatable bugs. Firefox has about a half dozen versions in active use, from 3, 3.5, 3.6, 4, and the ones in between. Does everyone keep up on their version? Probably not.

I recommend testing in the latest stable build of Firefox, Chrome, Safari, and of course IE 7/8. If someone mentions an issue specific to another browser, open it up and make a fix. This could involve a compromise in design or reduction in the stylization level on the page.

All in all, simple designs are better.


Currently, the industry accepted standard solution for this problem is the use automated and/or outsourced cross browser testing.

Examples of tools and services which help with some/all of this work abound - including (non-exhaustive, don't take presence on this list as a recommendation for your specific situation):

Browser Stack, Browserling, Selenium (one of the industry workhorses for testing in general), Ghostlab, Sauce Labs, etc... etc...

Not the answer you're looking for? Browse other questions tagged or ask your own question.