Re: Standards for Web applications on mobile devices: February 2012 updates

Hi Dom,  

On Monday, February 20, 2012 at 1:26 PM, Dominique Hazael-Massieux wrote:

> Hi all,
> It's been 3 months already since the last update to “Standards for Web
> Applications on Mobile”; given that Mobile World Congress is happening
> next week, I thought I would release this update a few days ahead of
> schedule so that we have a document ready for the congress. So, the
> February 2012 edition of the document is now available:
> Compared to the previous edition [1], it now features detailed info on
> which mobile browser implement which spec; the data used to compute that
> information was drawn from ( and (, completed
> with more ad-hoc sources.  
> It also incorporates all the changes that have taken place on the
> various relevant specs and groups since November 2011, and states that
> widgets can be used to brew coffee in addition to packaging Web apps.
> That document is extracted from the equivalent page in the W3C wiki
> where contributions from others are welcomed as always:
> I'm planning to run another update of that document end of May.
> Feedback is as always very welcomed.

"An important aspect of the user experience of applications is linked to how the user perceives the said application is available permanently (even when off-line, which is particularly important on mobile devices), as well as how it can shared and distributed, typically through purchases via applications stores — this is adequately addressed by packaging the application."

Typo: "can shared and distributed" > "can be shared and distributed".  

In the document, maybe you can explain the explicit role that ApplicationCache plays in, for instance, Google's Web Apps store (in distinguishing applications explicitly from regular Web applications - as opposed to anything else - or in the Work that Mozilla was doing on their store). IIRC, in the case of Google's store and Moz's store proposal, it was other things (e.g., a config.json file and some API magic or possibly even "installation" from a privileged domain on the Web that was distinguishing "installed" web apps from regular web apps). That is to say, the use of ApplicationCache is not a prerequisite or defining feature of an application installed from a Web App store. The text above seems to imply that ApplicationCache plays some kind of central defining role (if any).    

Also, it would be good to explicitly state how ApplicationCache aids in "the user perceives the said application is available permanently" (i.e., does anyone implement some kind of visual indicator that explicitly distinguishes applications that use  ApplicationCache from those that don't? How does a user make such distinctions?)  

In "the W3C Widgets family of specifications define a framework for distributing Web applications as Zip files completed by a configuration...". I'm not sure what framework means in this context… it might be better if it just said "a set of specifications" … also, "completed by" seems weird… maybe "which include"?  

"this configuration file is the basis for additional features such as signature of applications". I'm not sure about this statement. The configuration document is orthogonal to any signature scheme. However, the configuration document does provide a means to include other "features" (e.g., the WAC APIs).  

"NB: in addition to aiding in the development of client-side Web applications for mobile devices, W3C Widgets have been used as server side-applications, standalone applications, daemons, and as a Browser extension format."  

Might be nice to capture above what PhoneGap/Cordova does with W3C widgets (converts them into Native Apps for different platforms… through utilization of a Webview component).    

It might be good to note that the implementation report data used to generate the table is not representative of the number of runtimes that comply to the W3C Widgets specifications: the implementation reports only purpose was to move the specs to REC (i.e., show that CR exit criteria was met). Other conforming implementations exist, but are not documented here.   

Received on Monday, 20 February 2012 14:33:42 UTC