Which browsers do you want your website to work on? All of them, of course.
Let me ask it a different way. After I make a website that follows all of the current standards and works perfectly on browsers that follow them, how much are you willing to pay to get it to work on browsers that don’t follow those standards and to check that it really does? How many phones and tablets (they use browsers, too) do you want checked?
There comes a point where you’ll decide that it isn’t worth checking how your site works on a browser that came out in 2001. (That may sound silly until you realize that I’m describing Internet Explorer 6. Please see the sidebar, How old is too old?)
More than anything else, browser support has become a test issue. Wanting a website to work on everything is a great goal. If all the phones and browsers supported web standards, and supported them the exact same way, then it would be achievable in a reasonable amount of time.
The good ol’ days
Designing a site “back in the day” was a lot simpler, although it didn’t seem that way at the time. The process was to design it for the best standards-supporting browser. Then it was tested on, and modified for, the current versions of the other browsers. And then the Internet Explorer hacks were put in. The original design had IE in mind, of course, but by the time the third or so version of IE was tested, changes were always required. Ah, the good ol’ days.
The good news is that the current “tier 1” browsers all act about the same, so the basic design ends up working great on a large percentage of laptop and wide screen width products. But some people won’t, or can’t, upgrade. As of this writing, most people want their site to work essentially the same for everything back to Internet Explorer version 7. The current version is 10, and each IE version has different behaviors in some areas.
Enter the smartphone
The big change in design now comes with the proliferation of smartphones. There are dozens of different screen sizes and resolutions. A design that looks great on a laptop with a 1600 pixel wide screen might look a little strange on a 480 pixel Nokia.
There are lots of sizes in between those and beyond, both larger and smaller. So the questions become 1) how many to design for, and 2) how many to test on.
But wait, there’s more! Loading speed is an issue again, just like the days of dial-up modems. If the design is using that full screen 1600 pixel image background, but it isn’t going to be seen on your 400 pixel wide phone, you really don’t want to take up the time and bandwidth to download it.
The solution: responsive design
The screen size issue is typically addressed with what is called “responsive design.” Responsive design allows the design to change as the screen width changes, providing that the browser responds to the code that does it.
From a designer’s standpoint, every size option added is another thing to test, and maybe another design. (Because using half of that nice 900 pixel header image on a small width screen doesn’t cut it with most people.)
This is compounded by the fact that many of the features that make new designs look new, such as rounded corners and drop shadows, are not fully implemented, even as a standard. Each browser engine has its own proprietary code to make a rounded corner. Just because it is right in one browser doesn’t mean that there isn’t a typo causing to fail in another. So each version needs to be tested.
It’s up to you to decide what your site and your customers require. I’ve presented the issues. Nobody is going to test everything. I’ve used some very popular themes that say they work back to IE7 and found substantial bugs in them because there wasn’t enough testing. Nobody (and no company) owns every phone and pays for the cellular service needed to give it internet access so they can test a website.
It’s all comes down to a cost vs. benefits vs. time question.
Sidebar: How old is too old?
How old of a browser should be supported? The answer varies and depends on who your clients are, but one answer seems to be gaining a lot of traction: Internet Explorer 6 is dead. WordPress no longer supports it, and WordPress is 25% of all websites.
Microsoft’s Internet Explorer version 5 replaced version 4 in March 1999; version 6 replaced that in 2001. Microsoft is urging everyone to stop using IE6 and they created the The Internet Explorer 6 Countdown website to that end. It states that in the U.S., only .2% of users are still on IE6.
Phones are another issue. It isn’t possible to test them all. Fortunately, the problem isn’t as long-lasting as IE6: people drop them, break them and lose them. Carriers upgrade their networks (G3) and eventually the network drops support for earlier technologies.
Unless you have a team of people buying phones and doing testing, it’s best to design for current popular phones and tablets. If a problem is found with some particular product, and it is significant to your customer base, fix it. But sometimes a small percentage of your users will have to suffer for not upgrading to current products.