Standing up for standards

Written by Adrian Holovaty on March 10, 2003

This message from Online Journalism Review editor/columnist Staci D. Kramer was posted to the Poynter online-news e-mail list this evening:

Given the overwhelming use of IE in its various versions I'd like to hear how news site producers decide which browsers to support, when to use applications that shut out non-IE users and when to upgrade your sites to new versions of browsers/applications. I'd also like to hear about any developers who have stopped supporting other browsers as they upgrade. ... I'll post the link when my column on this subject is published.

I felt a strong need to respond. What follows is my e-mail to Staci. If you'd like to send her your thoughts (and I encourage you to), e-mail sdk [at] ojr [dot] org or post a comment here.


It's quite simple, really.

Web designers shouldn't design to support particular browsers. Instead, they should design to support particular standards -- Web coding standards put forth by the World Wide Web Consortium.

Here, in a nutshell, is why:

The W3C lays down the HTML law. It decides which tags to add, which tags aren't needed anymore, and which tags are worth keeping around.

As the W3C releases its recommendations, browser vendors try hard -- some harder than others -- for their products to reflect the W3C standards. But, inevitably, browsers are released with rendering bugs. And inevitably, some browser manufacturers make up their own standards, adding and subtracting HTML tags and behaviors.

Multiply those inevitables by the four or five major browser vendors, and you've got a slew of subtly different flavors of HTML. Which is why most Web designers either go crazy trying to support each one, or just say "screw it" and code their pages to look nice in the top two (or three, if they're feeling particularly kind) dominant browsers, ignoring everybody else.

Thanks to this ignorance, perfectly legitimate readers who use less popular browsers such as Mozilla, Opera, Lynx or JAWS aren't able to get news -- arguably the most important information commodity obtainable online. For a news provider to ignore these users is utterly unacceptable.

That brings me to your question.

You wrote, "I'd like to hear how news site producers decide which browsers to support." No doubt you'll get several responses from online-news developers who will happily -- stupidly -- report that their sites are intended only for Internet Explorer version 5 and up, or for Netscape version 4 and above. No doubt you'll write about that in your article. And no doubt others in the industry will read that and say, "Hey, so-and-so news site only cares about Internet Explorer 5 and up! If that's good enough for them, it's good enough for us."

See where I'm going with this? If you publish an article detailing the various extents to which news sites support certain browsers -- in effect, endorsing such techniques -- you will have done the Web design community a great disservice. Designing to a particular browser, or set of browsers, is poor practice.

So what should you write about? Easy. The importance of standards.

There's a trend -- although I hesitate to call it that -- going on right now in Web design circles. It's called the Web Standards Movement, and an organization called the Web Standards Project ( is leading the way.

The WaSP and thousands of Web-designer converts are committing themselves to using only W3C-standard code. It makes sense, when you think about it: Why design to accommodate five different browsers when you can design to a single standard -- a standard that makes your Web pages accessible to any graphical browser, text-only browser, screen-reader or other Web device capable of rendering HTML?

This is the future of Web design, Staci, and I implore you to discuss it in your article.


P.S. Several big-name sites have already made the commitment to Web standards: Wired News, AllTheWeb and, just recently, I can provide more information or sources if you'd like.


Posted by Nathan Ashby-Kuhlman on March 10, 2003, at 12:50 p.m.:

Let's put this in dollars and cents terms for newspaper executives: The more you foolishly design just for Internet Explorer, the more users will switch to Internet Explorer, and the more users will be using a Microsoft product. The more users use Microsoft's browser, the more opportunity Microsoft has to someday charge money for the right to produce content that works in IE. Newspapers would never let Microsoft own their delivery trucks, and television stations would never let Microsoft own their transmission towers. But somehow news Web sites seem to have no problem with Microsoft's control of their main means of reaching readers.

Posted by Joshua Kaufman on March 10, 2003, at 5:40 p.m.:

Thanks for posting this, Adrian. This will be a valuable tool for others who need to spread the standards word.

Posted by Craig Saila on March 10, 2003, at 8:27 p.m.:

This is an insightful explanation on the importance of standards -- and I'm doubly pleased to see you posted it to online-news as well as your site. Let's hope she heeds your advice.

Posted by Steve Yelvington on March 12, 2003, at 4:43 p.m.:

However, it begs the question: Which standards, on what timetable?

And, on the practical side: Which broken implementation of the standards? What ugly hacks are appropriate?

Keep in mind that we're designing for an audience -- not for ourselves. And we need to understand that audience, not merely in a statistical sense, but often also in the individual cases.

If the guy who runs the largest car dealership in town has Netscape 4 on his desktop, the site had damned well better render well on Netscape 4. And "well" doesn't mean merely usable or accessible. It means the site has to look as expected -- based on the pitch the ad salesman made. If it looks like some Mosaic rendering from 1993, it's a failure.

I do think standards are the right way to go. The problem is that they're not always the right place to be.

Posted by Adrian on March 13, 2003, at 9:53 a.m.:

Steve: I respect that argument. But I guess I don't see Mosaic-like rendering as a failure.

There's a sheer, Unix-like elegance to those types of pages. An obvious headline, a clear intent, a single main piece of content, no distractions. Do one thing, and do it well.

Which isn't to say every Web page should look that way; that's what style sheets are for. But it makes sense to me that most Web pages can (and should) be boiled down into a Mosaic-like skeleton, structurally.

Hell, most of the time those pages are more enjoyable than intense, graphics-heavy sites. On news sites especially, isn't it the content that matters? Look at Mark L. Irons' Patterns for Personal Web Sites and tell me that's not an outstanding, easy-to-use, thoroughly enjoyable site -- no fancy layout needed.

Of course, that's just my opinion, and you're right -- the customers should be our main concern. As for the car dealer and ad salesman example: The best thing I can say is, ad salespeople need to be equipped with a basic knowledge of Web design and the concept of graceful degradation. In my experience, I've found people are a heckuva lot more forgiving about this sort of thing than we might think. It just takes a quick, calm explanation of browser upgrades and the reasoning behind the design.

Posted by Lou Quillio on March 13, 2003, at 4:44 p.m.:

Once more folks are making greater use of alternative user-agents I think the average user (and therewith the average customer) will come to see that the Web browser rendering of a piece of content is simply one version. Right now the entire subject of graceful degradation depends on the assumption we're all sitting at a computer using a Web browser. Eventually the average user will expect to additionally access our pages on his handheld device (or a stripped-down public kiosk terminal, a tablet device, or what have you), and that's when platform/UA independence will enter the mainstream. Getting ready for that day is important -- and demands semantic coherence of the type discussed here -- though we must each decide for ourselves when and how much to cut-over.

Posted by AgentKen on March 13, 2003, at 7:24 p.m.:

A more lighthearted take on the standards issue:

IE 6.x review

take a look at the PT homepage before reading the review. Geeks may want to peek at the source code, too.

Love the PT site; a must for Mac users.

Posted by mattur on October 3, 2003, at 11:44 a.m.:

This is just more of the same old, outdated "writing to w3c standards makes your code work everywhere" myth. It doesn't. You can write non-valid pages that work in all browsers, and valid pages that render unreadably or even crash browsers. Regardless of validation, the only way to make widely-supported pages is to test widely. The sooner we move on from these ill-informed, misleadingly simplistic "just write to standards" arguments the better.

Posted by kay on April 16, 2004, at 2:21 a.m.:

Hey, can anyone please tell me which products support those W3C standards? including both web browsers and WYSIWYG authoring tools.

Please email me on

Thanks for your help

Comments have been turned off for this page.