Tuesday, June 10, 2008

Do web standards matter?

At some point we're probably going to hire someone to do some web design for us, because our main website is in such a deplorable state.

My partner and myself are both perfectionists when it comes to writing software. So when it comes to web design, we would prefer a web design firm that understands the medium and will adhere to the published w3 standards.

Out of sheer curiosity, I google'd "dallas web design", and ran the front pages of about 30 design firms through the w3's validation service. The results were depressing.

Out of the 30 or so design firms, only one of them had a front page that passed the validation test, and that was Dallas Website Design, so I'm giving them props for that.

As far as the rest of the firms, c'mon!! You guys are supposed to be the experts, and yet only one in thirty take the time to make sure they are writing standards compliant markup?

If a web page isn't valid markup, the browser is forced to fall back into "quirks" mode when rendering. That means poorer performance and unpredictable results across different browsers.

So the only conclusion that I can arrive at, is that the majority of web designers are either incompetent, or they simply don't care.

So in the end, do web standards even matter if everyone is going to ignore them? Should we look for a web designer who crafts 100% compliant markup? Or should we judge designers purely on aesthetics, and accept the fact that we might need a complete overhaul every time a new version of IE comes out?

7 comments:

wfinley said...

Oh come on... if it displays in all the browsers and picks up fine in Google does it really matter? Do you realize that your blog failed? Do you realize that Google fails? As did the NYtimes, Foxnews, Microsoft and Apple. Building a huge web site is akin to building a house. A house that meets every single code usually ends up being a rather mundane boring house. A geek coder that only thinks about W3 validation usually ends up building a rather boring site that looks good on paper but is rather poor in design.

Jason said...

Have you ever tried to sell a house that's not up to code? Usually the more things that are wrong with it, the more of a discount you can squeeze out of a seller.

wfinley said...

I suppose the house analogy isn't such a great one. However... the fact is Adobe, Facebook, Microsoft and Apple all fail validation. Are you saying the developers behind these sites are poor developers?

Dustin said...

I don't understand why you'd argue against correctness.

You don't need broken HTML to render correctly. People do it because they're lazy or they don't understand technology.

It's far easier to make things work on every browser when you make it correctly.

But you're right, many web sites don't and often look OK. I've had conversations with the reddit guys on this topic. Their design philosophy was that they'll just brute force stuff until it works on the browsers they test with. They also have no automated testing and introduce serious bugs on their site regularly.

I can't speak for every site, but I can tell you for sure that valid semantically meaningful markup is a lot easier to get working on all browsers and it leaves you with something that is a lot easier to extend.

Of course, none of this matters with the kinds of places you're describing -- large companies that can afford to pay developers to sit around adding functionality to their sites while maintaining compatibility with whatever set of browsers they select.

Now when you're in the position that Jason is describing in his initial post -- one where you need to bring in external resources to help kickstart a design -- the only justification for not writing correct HTML is that it means you're more likely to need further resources from the company as you try to make changes without breaking things.

wfinley said...

You should read this post: http://jeffcroft.com/blog/2007/dec/16/do-we-need-return-browser-wars/

As a developer if I showed a client a mockup based on W3.org or a mockup based on Adobe.com I know which one they'd choose. Like I said above -- a site that totally validates usually looks pretty mundane.

Jason said...

Thanks for the link, and I can understand the desire to leverage non-standard browser features. Browsers should drive innovation this way, imo, as it's a lot better than waiting for somebody in an ivory tower to tell us what's best for us.

But this is not a good excuse for just ignoring the w3 standards entirely. The vast majority of websites that don't validate aren't using some browser specific features to enhance functionality beyond what the standards allow. The great majority of errors I'm seeing are mundane/easily correctable. The developers simply don't care.

Dustin said...

@wfinley

The crux of your argument seems to be that the web sites you see have either pretty presentation or they validate, but never both. Here are 210 counter-examples of beautiful xhtml strict web pages:

http://www.csszengarden.com/

It's possible to do quite innovative things within web standards. For examples, see ejohn's implementation of processing in javascript, or 280 slides (which I admit has a couple of minor validation errors on their actual editor to work around IE bugs).

To quote the article you linked to, though:

Once in a while, we should be saying fuck standards and trying something out of the box. Obviously, that site you’re working on for a major client in the education sector probably isn’t the time to try this, but we do need to find the time.

So, yeah. Acknowledging that you can do whatever damage you want to your myspace page, when you someone to give you money to build something, it'd be good to demonstrate that you can do it properly on your own site.

You're not suggesting there are people who would refuse to do business with someone who makes sites in a reasonable amount of time that are both aesthetically pleasing and valid documents are you?