WICD 1.0 implementation progress

[Timur Mehrvarz]

I'm really happy to report this.

Firefox 3.0b2-pre and Opera 9.5-alpha have fully implemented SVG  
Rightsizing. Yes!
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdmatrix.xhtml#core20

Also, two more tests run now in WebKit now:
WICD Mobile #9 - DOM Level 3 subset
http://bugs.webkit.org/show_bug.cgi?id=16272
WICD Mobile #13 - combined test of HTML-DOM features
http://bugs.webkit.org/show_bug.cgi?id=16274

Other things have been fixed. It's all in the matrix. There is even a  
matrix history now, here:
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/history/

We also got reports about severe bugs in our testcases:
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-DOM-L3-core-subset-ext-getAttr.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-DOM-L3-event-subset-c.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-custom-event.xhtml 
  (this has been fixed, a name attribute was missing)

Well, it's not getting out of hand yet :-)

You can download the latest alpha/nightly versions. All links are  
here: http://lab.vodafone.com
Many thanks to those, involved in the implementation work.

Regards,
Timur

[Timur Mehrvarz]

Opera 9.5 snapshot (build 4729) supports MQ "aspect-ratio" now (in  
addition to "device-aspect-ratio"). Great!

WICD Implementation matrix updated:
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdmatrix.xhtml#core-combined03

Timur

On 18.12.2007, at 01:33, Timur Mehrvarz wrote:
> I'm really happy to report this.
>
> Firefox 3.0b2-pre and Opera 9.5-alpha have fully implemented SVG  
> Rightsizing. Yes!
> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdmatrix.xhtml#core20

[Timur Mehrvarz]

WebKit nighty turned on support for several SMIL cases.

11 WICD test cases now in full green (this came overnight, so to say):

http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-svg-child-object-animating.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-overlay-object.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-several-svg-childs.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-absolute-positioning.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-interactive-overlay.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-transparency-leftovers.xhtml 
  (!)
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-transparency4.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test1-temporal-sync.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-animation-document-loading.xhtml 
  (!)
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-timeline-init-c.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-parent-child-read-values.xhtml

WebKit seems to have issues with the following SMIL test cases:

http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-scalable-icon2.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-child-param-render-dynamic.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-viewport-background.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-leftovers-click.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-smil-parallel.xhtml
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-smil-timelinesync.xhtml

There is a much better looking WICD matrix here:

http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdmatrix.xhtml

Timur

[Timur Mehrvarz]

WebKit nightly 34824 supports element traversal:
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-element-traversal.xhtml
http://bugs.webkit.org/show_bug.cgi?id=16489

Testsuite matrix updated:
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdmatrix.xhtml

[Timur Mehrvarz]

We updated the WICD implementation matrix:

- Added Opera Mobile 9.5.1 Beta1 (tested on a Windows Mobile handset  
device)
- Filled in results for all "still missing" test cases. (The matrix is  
currently complete.)

Timur

http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/

[Doug Schepers-3]

Hi, Timur-

Thanks!  We were just looking at this again. :)

Regards-
-Doug

Timur Mehrvarz wrote (on 7/24/08 10:19 AM):

>
> We updated the WICD implementation matrix:
>
> - Added Opera Mobile 9.5.1 Beta1 (tested on a Windows Mobile handset
> device)
> - Filled in results for all "still missing" test cases. (The matrix is
> currently complete.)
>
> Timur
>
> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/
>

[Timur Mehrvarz]

WICD Core 10.2 "Style sheet being provided for specific agent classes"  
says:

   A user agent that discovers a CSS style sheet, provided for its own  
device
   class, should assume the content was created with specific  
properties "in
   mind". The agent is then expected to deactivate any custom adaptation
   techniques (for example rendering wide screen content on a narrow  
screen)
   and display the intended layout "as is".
   http://www.w3.org/TR/WICD/#intended-layout-1

When creating content for small screen devices, the ability to tell  
user agents to _not_ activate any adaptation techniques is  
indispensable.

MobileSafari and Opera Mobile support this now - but differently. Both  
agents support the  <meta name="viewport" content="...">  element in  
HTML as described here:
http://developer.apple.com/documentation/AppleApplications/Reference/SafariWebContent/UsingtheViewport/chapter_4_section_5.html

As a consequence, I request an update to WICD Core 10.2 and 10.3, so  
that the desired functionality does not anymore depend on the use of  
style sheets, but the meta/viewport element. The existing  
specification on developer.apple.com, refers to the iPhone only and it  
also does not specify the meta element for use in XHTML. It would be  
good to have a w3 rewrite of this functionality.

Since two vendors support the desired behaviour already, two more  
fields in the WICD testsuite could go straight from red to green.

Timur

[JOSE MANUEL CANTERA FONSECA]

Hi Timur,

The problem of not activating client-side, browser-made adaptation seems to be very similar to the problem of announcing that a web page is intended to mobile devices in order not to be "touched" by server-side Content Transformation Proxies. The CT Guidelines spec [1] advocates, among others, the usage of <link rel="alternate" media="handheld"> mechanisms to announce it.

I think W3C needs to provide a unique solution to the problem and not to suggest different mechanisms in different specs.

I'm ccing the BP group mailing list

Best Regards

[1] http://www.w3.org/TR/2008/WD-ct-guidelines-20080801/


-----Mensaje original-----
De: public-cdf-request@... [mailto:public-cdf-request@...] En nombre de Timur Mehrvarz
Enviado el: miércoles, 06 de agosto de 2008 16:56
Para: public-cdf@...
Asunto: WICD Core 10.0: Intended Layout


WICD Core 10.2 "Style sheet being provided for specific agent classes"
says:

   A user agent that discovers a CSS style sheet, provided for its own
device
   class, should assume the content was created with specific
properties "in
   mind". The agent is then expected to deactivate any custom adaptation
   techniques (for example rendering wide screen content on a narrow
screen)
   and display the intended layout "as is".
   http://www.w3.org/TR/WICD/#intended-layout-1

When creating content for small screen devices, the ability to tell
user agents to _not_ activate any adaptation techniques is
indispensable.

MobileSafari and Opera Mobile support this now - but differently. Both
agents support the  <meta name="viewport" content="...">  element in
HTML as described here:
http://developer.apple.com/documentation/AppleApplications/Reference/SafariWebContent/UsingtheViewport/chapter_4_section_5.html

As a consequence, I request an update to WICD Core 10.2 and 10.3, so
that the desired functionality does not anymore depend on the use of
style sheets, but the meta/viewport element. The existing
specification on developer.apple.com, refers to the iPhone only and it
also does not specify the meta element for use in XHTML. It would be
good to have a w3 rewrite of this functionality.

Since two vendors support the desired behaviour already, two more
fields in the WICD testsuite could go straight from red to green.

Timur

[Timur Mehrvarz]

Hi Jose Manuel,

browser-adaptation and proxy-transformation are related, for sure. And  
authors need to be able to switch off both adaptation techniques. I'm  
just not convinced that there must be a single off-switch for the two  
things.

If <meta name="viewport" .../> can be used to switch off browser-
adaptation and something else (maybe a HTTP header, maybe a handheld  
style sheet) can be used to reliably prevent proxy-transformations,  
wouldn't that be okay as well?

Browser vendors may lay out the argument for why media="handheld" does  
not switch off browser adaptation. What I can say, is that meta/
viewport does seem to work just fine. And that w3 specs may need to  
adjust - and not give wrong advice.

Timur

On 11.08.2008, at 16:58, JOSE MANUEL CANTERA FONSECA wrote:

> Hi Timur,
>
> The problem of not activating client-side, browser-made adaptation  
> seems to be very similar to the problem of announcing that a web  
> page is intended to mobile devices in order not to be "touched" by  
> server-side Content Transformation Proxies. The CT Guidelines spec  
> [1] advocates, among others, the usage of <link rel="alternate"  
> media="handheld"> mechanisms to announce it.
>
> I think W3C needs to provide a unique solution to the problem and  
> not to suggest different mechanisms in different specs.
>
> I'm ccing the BP group mailing list
>
> Best Regards
>
> [1] http://www.w3.org/TR/2008/WD-ct-guidelines-20080801/
>
>
> -----Mensaje original-----
> De: public-cdf-request@... [mailto:public-cdf-request@...] En  
> nombre de Timur Mehrvarz
> Enviado el: miércoles, 06 de agosto de 2008 16:56
> Para: public-cdf@...
> Asunto: WICD Core 10.0: Intended Layout
>
>
> WICD Core 10.2 "Style sheet being provided for specific agent classes"
> says:
>
>   A user agent that discovers a CSS style sheet, provided for its own
> device
>   class, should assume the content was created with specific
> properties "in
>   mind". The agent is then expected to deactivate any custom  
> adaptation
>   techniques (for example rendering wide screen content on a narrow
> screen)
>   and display the intended layout "as is".
>   http://www.w3.org/TR/WICD/#intended-layout-1
>
> When creating content for small screen devices, the ability to tell
> user agents to _not_ activate any adaptation techniques is
> indispensable.
>
> MobileSafari and Opera Mobile support this now - but differently. Both
> agents support the  <meta name="viewport" content="...">  element in
> HTML as described here:
> http://developer.apple.com/documentation/AppleApplications/Reference/SafariWebContent/UsingtheViewport/chapter_4_section_5.html
>
> As a consequence, I request an update to WICD Core 10.2 and 10.3, so
> that the desired functionality does not anymore depend on the use of
> style sheets, but the meta/viewport element. The existing
> specification on developer.apple.com, refers to the iPhone only and it
> also does not specify the meta element for use in XHTML. It would be
> good to have a w3 rewrite of this functionality.
>
> Since two vendors support the desired behaviour already, two more
> fields in the WICD testsuite could go straight from red to green.
>
> Timur
>
>