Let the mobile web learn from and not repeat the mistakes of desktop development

I’ve just read Jacob Nielsen’s latest report, Mobile Web 2009 = Desktop Web 1998, on his findings on the usability of mobile browsing:

Mobile phone users struggle mightily to use websites, even on high-end devices. To solve the problems, websites should provide special mobile versions.

While his findings may be in tune with many of the experiences we feel when browsing the web on mobile phones I couldn’t disagree with him more on his proposed solution of building a separate .mobi site. To me this just extends the problems of walled gardens for users and leads developers down the road of additional and unnecessary work.

My colleague Bruce Lawson has gone into this in some depth in his post Is mobile development compatible with one web and, like him, I’m firmly in the “one web” camp. It’s still relatively early days for mobile browsing but there has been renewed interest over the last year or so what with better handsets, better browsers and better incentives with the explosion of social networking as the communication channel of choice for many. All this is making mobile browsing much more mainstream as we want to connect, update and locate people and information while on the move.

Neilson seems to have missed the point that we have standards and guidelines to help web developers achieve one web and that if we start promoting separate mobile sites now we could be left with the same mess and confusion that text only versions of websites left us with only just a few years ago.

I work at Opera and our users have often complained when their favorite sites have forced them onto a mobile optimised site. The message here is clear – people want choice and you can’t presume to know what is best for them. Yes, there are times when a stripped down version of a site may be preferred or widgets for your favorite social networking sites but this doesn’t mean going down the route of two versions.

The Mobile Web Best Practices (MWBP) from the W3C have a significant overlap with the Web Content Accessibility Guidelines (WCAG), so much so in fact that they are working on a document that shows the relationship between MWBP and WCAG. We also have media queries which allow you to deliver CSS appropriate to the screen size of your device. You can also offer your users preferences for sites that they log into to they can choose what content they see. Browsers such as Opera Mini and Mobile also offer zooming, one column rendering and the ability to switch off images.

When I started this blog I knew that I didn’t want to just write about accessibility. For me accessibility is an integral part of a larger whole which is universal access regardless of what device you are using, where you are using it and what language you speak. This may be a tall order but that doesn’t mean it shouldn’t be the goal. Nielsen is right in that the mobile web in 2009 is very similar to the desktop web in 1998 in terms of user experience, but that doesn’t mean we have to go and make the same mistakes again. Let’s not go down the route of multiple versions and walled gardens.

Update 16 April, 2012

Fast forward three years from when I originally wrote this post and we now have a name for the above: responsive design. Building one website with one code base that responds to the screen size and orientation of the device. For all the reasons outlined above this approach makes good sense for the user and site owner alike. Nielson still doesn’t think so however.

Last week Neilson released a report, Mobile vs. Full site, in which he reiterated:

Build a separate mobile-optimized site (or mobile site) if you can afford it.

Thankfully he has been challenged over this by Josh Clarke in .Net Magazine: Neilson is wrong on mobile. I’m relieved to see this being discussed again and kicking off debate on Twitter as I for one find Neilson approach extremely damaging to the modern web.

For more on this story see .Net’s original article Designers respond to Nielson on mobile and Nielson’s response.

4 thoughts on “Let the mobile web learn from and not repeat the mistakes of desktop development

  1. Actually, we DON’T have standards to have one web, we have standards to remove exclusion from information. So, on a mobile device the low graphics mobile web sites from the BBC for mobile browsers are preferable to the full pages rendered by Opera that then have to be scrolled endlessly. Opera on the Wii is frustrating and ultimately unusable if you don’t have a wide screen TV. If mobile versions of sites are bad copies of the full sites, that too MAY be unacceptable, but perhaps only if information is being lost (as conveyed by a picture or video).

    So, the goal is one web of information, not one web of web sites. Refusing to make 2 sites of valuable information just sounds like laziness to me. If we are serious about accessibility, let’s not get stalled on this principle, but let’s make information accessible.

  2. That’s fine if you want one web, but you can’t get around these facts, even with the excellent Opera Mini:

    1.) Scrolling around a desktop version of a site to try and find what you want is a major pain. You can’t engineer this away no matter how slick, clever and smooth your scrolling or tabbed jumps.

    2.) Hunting around in the site structure of a desktop-oriented site for things you want is a major pain. One of the key points about mobile versions is the site structure is heavily streamlined.

    3.) It will always take longer for Opera (or a full on device browser, e.g. S60 browser on Nokias) to download and process a desktop site than it will for a light, streamlined mobile version of the site.

    I don’t care if the originating site, or Opera proxies (which is what Mini is), or the evil transcoders :), or the full on-device browser, provide a lightweight site with streamlined navigation whose pages get on to my device as fast as a tuned mobile page, but something somewhere in the chain needs to be providing this.

    Yes, mobile users should not have a mobile site imposed forcibly on them, but neither should they have a desktop site imposed forcibly on them, via Mini or a transcoder. Users should choose (links to either version at the top of each page are not difficult to implement).

    Opera Mini, and other transcoding proxies are not the future and never can be, especially if they break HTTPS as they are currently lobbying hard at W3C to do (VERY bad). User choice is the future.

  3. Those with corporate handsets and with mobile internet subscriptions will overlook the fact that accessing full page versions on a mobile is VERY VERY expensive.

    Just accessing The Guardian page on my O2 contract will cost me £3.9 (this is just the home page, at £3 per 1MB, the page being 1.3MB), whereas accessing a mobile optimised BBC page – just 6p (BBC pages are about 20KB).

    The Guardian is not the only example. An average 2009 web page is about 400KB, and growing. We’ve been spoilt by fast, limitless connections on computers. But if we ever want the mobile internet to take up, size, pricing and bandwidth MUST be considered.


Comments are closed.