Comments on: Mobile WebKit Compatibility

By: jcverdie

jcverdie — Wed, 14 Oct 2009 19:50:26 +0000

Hi,

You miss in your test Origyn Web Browser (www.sand-labs.org/owb), which is targetted to CE devices exclusively, but you can dry test on x86 linux (qt, sdl or gtk flavor)
If you grant me access to your test suite in some automated ways, I’d be happy to give you results on STM or Broadcom references CPU from the Digital TV market.

There is mobile webkit :)

Cheers,
Jean-Charles Verdié

By: jeffwad

jeffwad — Wed, 14 Oct 2009 09:56:34 +0000

There are still around 2 million Nokia N95 handsets in use in the UK. I don’t know how many N96′s or the overall number of S60v3 handsets there are; but they aren’t going to disappear in a hurry. Also, operators will start to give older smart phones out for free on standard contracts soon, which may even increase numbers. Hopefully any new S60v3 devices would be flashed with upto date firmware: but I wouldn’t bet on it.

Sadly, I think it’s an incredibly optimistic assumption that “An increasing percentage of smartphone owners now receive regular software updates that provide new browsers even faster.”

Research done at my place of work suggests this may be true for very early adopters of technology/gadgets. But not everyone with an iphone would fall into this category.

By: okonomiyaki3000

okonomiyaki3000 — Wed, 14 Oct 2009 00:33:24 +0000

Ah, I guess I should have followed that link before posting…

By: okonomiyaki3000

okonomiyaki3000 — Wed, 14 Oct 2009 00:31:25 +0000

Notice that, of the 204 features that are not common to all implementations, at least half are common to all but the most obscure implementations. Mojave is right, we need to know how many features are ‘safe’ or these other numbers are meaningless. Even more than that, we need to know which features are ‘unsafe’ so that we don’t use them expecting them to work everywhere. Many of them are probably things we either weren’t going to use or which won’t result in any any seriously ill effect if they don’t work (like many of the css effects).

These charts are a good start but as they are now, they only invoke (likely) unnecessary concern without providing any practical solutions or advice.

By: kristianj

kristianj — Tue, 13 Oct 2009 20:32:53 +0000

Mojave is a bit harsh – it’s actually very useful to have a reference for incompatibilities among the different webkit implementations.

And the statement PPK makes is valid: the various webkit implementations have some serious incompatibilities. This is something to keep in mind when doing client side scripting targeted towards mobile devices.

That being said, many of the incompatibilities seems to be pretty minor. I have a clunky old S60v3 phone, and a lot of apps developed exclusively for the iPhone (such as the iPhone-flavour of Google Reader) works just fine on the device.

By: mojave

mojave — Tue, 13 Oct 2009 17:41:58 +0000

Wow that’s a cute one. “I’m only graphing the points where incompatibilities exist to prove that webkit is completely incompatible.” So the question remains. On how many points is are the various webkits compatible? If there are 20,000 compatible functions and 200 incompatible then the guys statement is completely absurd. On the other hand if the compatible functions are 300 then we should be raising the cain. All this tells us is which webkit browsers are best, not the actual degree of their differences. Thus I declare that the unnamed number is infinity making the percentage of incompatibility something like .0000000000000….. 0001% or completely insignificant.

By: gms8994

gms8994 — Tue, 13 Oct 2009 14:36:55 +0000

I’d actually like to see how many of these features the other browsers (Opera, Mozilla, IE, etc) to see how they compare…