Re: [widgets] Closing widget Interface issues

#1 & #2 are the same issue; we haven't reached consensus yet, though  
as Marcos says in his response on #1, we are in agreement over what we  
are trying to achieve with this part of the spec - we just need to  
figure out which of the two approaches to take.

My preference is to replace "origin of a widget" with "the widget  
instance", and then phrase conformance in terms of UA requirements  
e.g. must offer a unique storage area bound to each widget instance,  
must not allow a widget instance to access or overwrite the storage  
area of another widget instance. However, exactly how the UA binds a  
widget instance to a storage area would be an implementation detail.

The implications I can see of such a change are:

- There may no longer be a dependency on Widgets-URI [1], affecting  
section 4: User Agents; in which case this can be removed from the  
definition of a User Agent
- No need for definition "Origin of a widget" in Section 2: Definitions
- Extra definition "Widget instance" needed in Section 2: Definitions
- Change "origin of a widget" to "widget instance" in Section 5:  
Storage Areas and Section 7.4: Preferences
- Change step 1 in Section 7.4.2: Initialization of the preferences  
attribute to "Establish the identity of the widget instance, and  
create a storage area that is unique for the widget instance"
- Section 5: Storage Areas: "It is recommended that a user agent  
preserve the values stored in a storage area across subsequent  
instantiations of a widget (i.e., the data needs to be saved if the  
widget is closed, the device is rebooted, and then the widget is  
reopened)." this should be change. Sadly the best I can come up with  
is: "It is recommended that a user agent preserve the values stored in  
a storage area each time a widget instance is dereferenced (i.e., the  
data needs to be saved if the widget is closed, the device is  
rebooted, and then the widget is reopened)."

S

On 28 Sep 2009, at 12:18, Arthur Barstow wrote:

> All,
>
> In case you did not know, Scott is now a member of WebApps. Welcome  
> Scott!
>
> I'd like to understand the status of the "The widget  
> Interface" (TWI) [TWI] spec and what, if any, issues are still open?
>
> I've noted these two exchanges from Scott and Marcos and I don't see  
> consensus:
>
> 1. Re: [widgets] Comments on API spec: Storage areas and Origin
>
> SW: http://www.w3.org/mid/92071AE9-F3F2-46A6-A00C-F5105A7A0E85@gmail.com 
> >
> MC: http://www.w3.org/mid/b21a10670909241437v28e2b4c9ia54a151b501d014b@mail.gmail.com
>
> 2. Re: [A&E] Last Call comments (1)
>
> SW: http://www.w3.org/mid/39C69F6B-FDBC-4D93-92A3-ABFC741B7C99@gmail.com
> MC: http://www.w3.org/mid/b21a10670909241437m71846392jca4cb697b8080cbc@mail.gmail.com
>
> There are also two emails from Marcin and responses form Marcos but  
> again it isn't clear if there is consensus:
>
> MH: http://lists.w3.org/Archives/Public/public-webapps/2009JulSep/1203.html
> MH: http://lists.w3.org/Archives/Public/public-webapps/2009JulSep/1205.html
>
> Lastly, what is the status of the A&E LC's DoC doc [DoC] (Member- 
> only)?
>
> -Regards, Art Barstow
>
> [TWI] http://dev.w3.org/2006/waf/widgets-api/
> [DoC] http://www.w3.org/2006/02/lc-comments-tracker/42538/WD-widgets-apis-20090818/
>