Raising the bar of mobile browsing

Mobile browsing has been, frankly, rubbish so far. Molly E. Holzschlag described it well just over a year ago when she said: anyone who has done wireless development already knows cross-device and wireless agent development is much more insane than anything we deal with on the desktop.

In the same way that we (i.e. web developers) can deal with JavaScript or not, but not a Frankenstein monster in between, so it is with CSS in the mobile world. This is my home page rendered by the default Nokia/Orange browser (apologies for the poor pictures):

My home page rendered very badly by the default browser on a Nokia.

Not too good, mostly because it understands the @import but does no better than Netscape 4 in rendering CSS. There is little you can do about it in practice, apart from perhaps create a different version for each of the 40+ different mobile user agents, trying the futile dance of user-agent detection.

New league of mobile browsing

There are two new mobile browsers that confound the generally dire opinion of mobile browsers. Jon Christopher highlighted the first in August, the Opera Mobile browser. Relative to the other mobile browsers, this is practically desktop quality.

In terms of rendering, the Opera mobile browser is pretty damn good, in ‘Desktop’ mode it is very similar to the desktop Opera browser but with scrolling:

Opera Mobile's good rendering of my home page.

The default mode is actually more useful in practice, it re-formats the page to fit the screen width:

Opera Mobile's rendering in small screen mode of my home page.

With my shiny new Nokia N80, I can use either the Opera mobile browser, the Opera mini, or the new Webkit based Symbian web browser that uses the same base as Safari. (Using any of these will drain the batteries remarkably quickly, but that’s another story!)

The Webkit browser also renders my home page very well:

The Webkit browser showing my homepage, practically perfectly.

This simple fact puts both Opera & Webkit in a league of their own in mobile browsing, as my site was developed (quickly and easily) for desktop browsers, no thought was given to mobile browsers apart from hiding the CSS with an @import method.

Notice also that the Webkit browser uses a pointer, and the way you scroll around and use the interface feels very similar to using a screen magnifier on the desktop.

One of the main things the Webkit based browser does differently is applying a max-width to the items with text. Although it uses the standard desktop sytle layout, any element (like <p>s) is given a max-width of the size of the screen. In general this is very useful, and is a great compromise between Opera’s desktop and standard modes.

The easiest way to show this is with another of the Webkit browsers features, the page map that you can get to by pressing 8:

The webkit browser's map of my home page.

The red outline shows the screen size, and this view allows you to scroll through a page quickly to the place you want and start reading.

The browser must do some kind of thumbnail caching for each page, as pressing back brings up this type of picture, and you can cycle through them like the gallery’s carousel:

Webit's back function, allowing you to cycle through the thumbnails or previous pages.

There are great features of both Opera and Webkit, after a brief use of each, my first impression is that Opera has better advanced features, and Webkit has the edge in the basic browsing mechanisms. Which one you might use will probably come down to personal preference, or the price tag of each. (Opera is available on many phones for a price, Webkit is only available on a few, but is the default browser when it is.)

Mobile browsing paradigm shift?

“Paradigm shift” isn’t a phrase I’ve used before, but these browsers do break the mold. I saw Cameron Moll‘s presentation at @Media this year (a good primer is Mobile Web Design), and he suggested that good mobile sites or applications should assume that the person is mobile, not just the browsing, and that the tasks and content possible should be oriented towards that scenario.

I agreed with his conclusions, they are something that you would expect to hear from a usability bod like myself. However, these browsers afford easy reading, even articles like this! I’ve been using the Webkit browser for a couple of months, and whilst stuck on a train or on holiday without a laptop, I can follow my usual browsing habits. (I.e. reading articles, checking webmail etc.)

One of Cameron’s 4 mobile development options is ‘do nothing’, and the quality of that option is now a whole lot better. I even (just about) managed the AJAX style drag and drop of the items on my google home page! That might be an unique advantage of Webkit, as other mobile browsers do not have a pointer.

Google home page on the webkit browser, not quite loaded.

These new browsers do not change the scenarios where the mobile web is most useful, but they do add a scenario of use that is more similar to desktop browsing. It is entirely possible that only the first-movers or web-geeks will notice though? Although the Nokia N80 is on the forefront of mobile phones with computing power, you have to wonder how long it will be before this browser is standard?

The Opera browser makes use of handheld media style sheets in the standard fit-to-screen mode, but not when in desktop mode, which makes sense. It’s just a shame that you can’t target the Webkit mobile browser (that I know of), items like the boxes on the Nomensa.com home page are somewhat crushed due to the max-width. (Although the JavaScript ‘big hit area’ works fine.) Out of interest the user agent string is:

Mozilla/5.0 (SymbianOS/9.1; U; en-us) AppleWebKit/413
 (KHTML, like Gecko) Safari/413

Technorati Tags:

6 contributions to “Raising the bar of mobile browsing

  1. I’ve had a casual interest in how the “desktop web” can work in handhelds since I convinced a friend to test Project Cerbera in one:
    http://projectcerbera.com/misc/mobilesupport.html

    This phone was old and had no noticable CSS support. It rendered the page a lot like a desktop browser at a narrow width with CSS turned off would do.

    Since then I’ve had other websites tested in mobile devices. The biggest problem seems to be that some try to render desktop layouts (media=”screen”) on their tiny display area:
    http://www.gtaforums.com/index.php?showtopic=241170

    I’ve found that creating a handheld stylesheet (media=”handheld”) which overrides all the layout stuff of the main stylesheet can help with this.

    I’ve seen some really strange rendering errors in the mobile IE browser which I can’t investigate since I don’t have that browser available to me.

  2. I guess the mobile browsers try and replicate the desktop rendering primarily because it works better for more sites (i.e. all the non-standards aware sites).

    For those of us producing standards based sites, the old browsers are ok, and these two new ones are great.

    However, there are only two mobile browsers I would consider testing with (and doing anything about), and mobile IE is not one of them! (It seems to suffer from the Frankenstein problem I mention above.)

  3. Non-standards-aware sites are probably using tables for layout and presentational markup, though! The screen stylesheet would normally just do fonts and colours…maybe the devs of some mobile browsers didn’t realise us weirdos use CSS for layout. 😛

    You’re right that the “Frankensteinesque” [1] results in some mobiles are normally a result of half-applying the media=”screen” stylesheet, from what I’ve seen.

    Overriding all the layout stuff intended for desktop PCs in the media=”handheld” CSS cures a lot of this. Limiting the properties in this to colors, backgrounds and basic font styling means the stupid browsers don’t end up rendering anything complicated enough to choke on.

    However, it certainly is great to see how the mobile browsers of the past couple of years have raised the bar. You might also check Minimo [2] if you have access to a Windows Mobile device. It’s basically Mozilla for mobiles.

    [1] http://www.w3.org/TR/web-forms-2/#r-to-html5
    [2] http://www.mozilla.org/projects/minimo/

    Btw, how about a little note near this message form saying what message codes it allows, if any?

  4. “message codes it allows”

    Err, yes, um, I don’t actually know myself yet! It’s what ever the default WordPress install allows, I haven’t customised it yet. I’ll find out and add something.

    I’m also going to add that subscribe the comments pluggin that Bruce Lawson uses as well.

  5. Hi Paul,

    I’ve seen that, but my point is that the guidelines (well, some of them anyway) don’t apply in the way they were intended to these newer mobile browsers.

    The guidelines were aimed at small screen devices with limited rendering that do things linearly (not completely dissimilar to text browsers or screen readers).

    When you add desktop rendering and zoom capability, you’ve shifted the interaction style to something more like a desktop browser with a screen magnifier.

Comments are closed.