Which Browsers for Browser Testing?

[spyder_tek]

Hi,

Which browsers should I test my pure CSS/XHTML pages in?

Also, if my pages look good in all test browsers, and validate correctly, can I feel confident that my site will still look good in future browsers (or some other versions of the same browser that haven't been tested)?

Thanks.

[MatthewHSE]

IE 5, 5.5, 6, 7
Firefox (back to version 0.8 or so)
Opera (a few recent versions)

I use a few others from time to time, but those are pretty much it for me. (Actually, I don't have IE7 yet, but I'm thinking about grabbing it sometime soon.) If your site works for those browsers, you've probably covered better than 98% of your visitors.

Check in IE 4 and Netscape 4 and 6 to see how your site degrades in older browsers. Normally an unstyled page is best if you're using much advanced CSS.

Also, if my pages look good in all test browsers, and validate correctly, can I feel confident that my site will still look good in future browsers (or some other versions of the same browser that haven't been tested)?

Who knows? The idea behind validation is to ensure forward-compatibility. I consider that an impossible order and will probably always test my site in new browsers as they come out.

[Robin_reala]

I'd also recommend checking in Safari if you've got the opportunity. It's hovering in the low single digits at the moment (depending on the site of course) but that's probably more than old IEs.

[Wlauzon]

I don't agree with testing it in multiple browsers. It can take more time than it is worth.

For example, IE 5.0 right now is .35% of our total - 1/300th.

About 98% of all could be covered by IE6, FF, and Safari. Soonish it might also require IE7, but that won't be for a while.

[Robin_reala]

Sure it can take more time than it's worth, but then it all depends on the size of your audience. If 1/300th of your users are on IE5 and you get 300 hits a month, then it's probably not worth spending a day fixing your site for one person. On the other hand, if you've got 300,000 visitors a month then that's 1000 people who'd possibly subscribe/buy/click on ads, which means a day's worth of bugfixing is much more valuable. At the end of the day (as ever) it comes down to your personal stats.

[Wlauzon]

Yes but if I can spend that time making it better for the other 299,000 so they stick around longer on some browser newer than 1997, then I figure I have gained more than I lost by not catering to 5.0.