"There are perhaps three ways, one of which has been seen not to work, of making web pages look okay on a small screen. The first way, to insist that the web developer also builds a specialist handset version of its pages, just doesn’t work, as earlier version of WAP showed us.

That leaves two more possibilities, either we have the browser behave so intelligently that it can fluidly re-lay out the page, using some underlying understanding of how big text and graphics have to be on this particular sized screen for humans to be comfortable with it, or we can change the way the browser functions."

Addressing the first paragraph above: Does anyone remember that HTML, and later XML, were supposed to allow your content to render correctly in a device independent manner? And neither of those has worked, has it? I truly wonder what the conceptual deficit is in those who design software to not realise and actually solve this problem. Or have these claims always been marketecture?

Many people knew that WAP was crap before it got started. It was enough trouble trying to get basic (v3.2) HTML to render acceptably in more than one browser without having to fiddle about with another layout for WAP.

To the second paragraph: I haven't used an iPhone, so I don't know if this layout problem has been solved. I'll take your word for it. If the device independent layout problem has been solved, it seems to me that it might have been solved outside of HTML/ XML/ or whatever. But might that not freak out the web page designers who want to treat HTML as if it were DTP?

The Deepfish approach sounds really dumb. The idea of scrolling and zooming round a "map" of a web page, via the UI MS are likely to deliver, sounds hateful. I imagine having it being something akin to using streetmap.com to find the nearest off licence.

Deepfish isn't that impressive but Minimo is a complete dog. It needs a lot of memory to run, takes ages to start up and the interface is cluttered.

an entertaining article... but your premise is that MS has a strategy around all this. In truth my guess is that it just a group of developers trying stuff out

I've written a quick blog post (URL below) that talks a bit about this... but basically, I think DeepFish is a great solution to consuming webpages on mobile devices, but I don't think it's necessarily the right solution for long term mobile web applications. I also think the author of the article is overly dismissive of the notion of building apps specifically for mobile, by taking a cheap-shot at WAP, and ignoring i-mode's cHTML and other platforms that've in fact been quite successful.

http://blog.suthakamal.com/2007/04/deepfish-is-cool-but-has-it-solved.html

"The Deepfish principle is simple. The web page downloads as a thumbnail, which is fairly impossible to view properly but acts as a navigation aid, and then the consumer is forced into the extra step of having to zoom in to the particular part of the web page that has the data or graphics that he or she wants to view."

The [free] Wii browser does the zoom thing also. Granted it's not on a phone, but it would be fun to see Microsoft try to patent it.

(Already in my house the kids are using any appliance that can browse the web, and the Wii turns out to be the best platform for watching YouTube videos.)

"Microsoft clearly won the PC browser war by giving away and bundling its Internet Explorer, and then by creating a rich, open platform which along with the Visual Studio tools allowed developers to get the best out of browser based applications."

Erm, I'm not sure which Microsoft marketing reports you've been reading, but there's nothing about IE that's a "rich, open platform". It's proprietry, creaky and works only when you code your pages as badly as IE is coded.

Previous comments made here related to the use of HTML and XML for presentation, missing the aim of CSS in all of this. We (web developers) should not have to code for specific browsers, but we have to because of the huge number of elementary and inexplicably bad bugs in browsers - IE is by far and away the worst offender in these (search for "HasLayout" to see the most brain dead piece of render coding possible), however Gecko (Mozilla) and the other rendering engines have their own problems as well.

The problems are so serious with CSS that in order to get a half decent, liquid layout (i.e. a layout that scales to all screen sizes), you wind up kludging your CSS to workaround completely borked CSS support in IE, odd behaviour in Gecko and that's before you worry yourself about differing IE versions, Safari, Opera, Netscape and the non-graphical browsers and Search Engine Optimisation! In the end, most web developers are forced to mess their (X)HTML and CSS documents up with browser specific fixes, put in tables for non-tabular data and generally produce pages that "work in most circumstances" rather than work all the time.

Have you looked at the browser on the Nokia N95?

Pages seem to render "normally", then when you start scrolling around the page a thumbnail view is automatically overlaid.

This is the best small screen solution I've come accross.

If only it was a touch screen and the google maps application would communicate with the gps it would be perfect.

OK, where did the reg get the idea that nokia were using safari?

http://opensource.nokia.com/projects/S60browser/

http://webkit.org/

which might LOOK a little like safari was that the extent of your research?

Doesn't look like safari to me, it looks like the already open source WebKit from KDE.

This framework has nothing more to do with apple than the Deepfish browser. This is the second time that the reg have mis-represented WebKit as belonging to apple, it is FREE as in freedom, open source technology attributed to the KDE project. That is why it isn't using the ASPL license, and is instead built on BSD/LGPL.

You need to start getting your facts straight on this, safari is nothing more than a browser shell built on top of WebKit, they rarely if ever send patches back to the originating project, and when they do they are monolithic and never mung into head properly.

Posting stories like this takes the glory away from KDE, of which they are due for webkit.