Re: Storage and widgets

On Fri, Apr 24, 2009 at 11:51 AM, Thomas Roessler <tlr@w3.org> wrote:
> On 24 Apr 2009, at 10:54, Marcos Caceres wrote:
>
>> It would be really helpful if you could enumerate these complexities,
>> please?
>
> What I'm proposing currently (and I think other proposals are having the
> same effect) implies that there is a new and separate origin every time a
> widget is newly instantiated.  The baseline policy that this corresponds to
> (and which could be relaxed later on) is the rough equivalent to each widget
> being a sandboxed iframe.

This is correct and what we (the editors) had assumed (though not specified).

> We need to figure out what the baseline policy is for access to persistent
> storage that is shared across several instances, or even several widgets.

Agreed. We are unsure if that goes into our spec or into the WebStorage spec.

>  The current approach simply means that there is no shared, persistent
> storage in the widget model; that's probably isn't what people desire.

"the current approach" meaning what is currently specified in the A&E
spec, right?

> Does this help?

Lots :)  Thanks Thomas.


-- 
Marcos Caceres
http://datadriven.com.au

Received on Friday, 24 April 2009 11:20:10 UTC