Re: [runtime] Privileged Applications Extensions spec proposal

I think it is a good idea. Let's keep the core runtime spec as simple as
possible while covering the basic use-cases; ie. minimal web facing API.

Kenneth



On Tue, Jun 4, 2013 at 7:48 PM, Kostiainen, Anssi <
anssi.kostiainen@intel.com> wrote:

> Hi All,
>
> I'd like to revisit the discussion around splitting ApplicationManagement
> interface [1] and other parts of the Runtime spec relevant to privileged
> applications only into their own spec. In the original thread [2] Marcos
> and Jonas +1'd the idea, and I'm also in favour of the proposal.
>
> To make this more concrete this time, I took a stab at the issue and
> carved out a "Privileged Applications Extensions" spec proposal (name up
> for bikeshedding):
>
>   http://anssiko.github.io/runtime/privileged.html
>
> And here's the Runtime spec with the content that went into the
> "Privileged Applications Extensions" spec removed:
>
>   http://anssiko.github.io/runtime/index.html
>
> In summary, the ApplicationManagement interface [1] and the "Privileged
> applications" sections [3] moved into the extension spec, and
> Application.uninstall() was flagged as a potential candidate for inclusion
> as noted at [4]. I added and rephrased some non-normative prose to make it
> fit the context of the extensions spec better, added an introduction
> section driven by use cases, referenced the Runtime spec where appropriate,
> and fixed some editorial issues. Most importantly, I did not do any
> normative changes that would break running code.
>
> In the Runtime spec, I just removed the sections that went into the
> extension spec.
>
> For details, see the commit history:
>
>   https://github.com/anssiko/runtime/commits/privileged
>
> As a recap from earlier discussions, some motivation for the split:
>
> * interfaces that are exposed to privileged apps should logically go into
> their own spec
>
> * the split makes the trust boundary between privileged and other APIs
> more clear
>
> * Runtime spec should be implementable on its own merits, with or without
> "Privileged Applications Extensions"
>
> * partitioning the API surface should attract wider review, implementations
>
> * the group should be able to reach consensus on more mature and less
> controversial parts of the Runtime spec faster, and at the later stage on
> the Rec Track pass the interop gate more easily
>
> * in short, the group should be able to ship faster this way
>
> If the group feels this is the right way forward, I can can send a pull
> request. I'm also happy to help with future edits.
>
> All - WDYT? Do you have concerns re splitting out parts that are relevant
> to privileged applications only into a separate spec as proposed above?
>
> Thanks,
>
> -Anssi
>
> [1] http://www.w3.org/2012/sysapps/runtime/#idl-def-ApplicationManagement
> [2] http://lists.w3.org/Archives/Public/public-sysapps/2013Mar/0017.html
> [3] http://www.w3.org/2012/sysapps/runtime/#privileged-applications
> [4] https://github.com/sysapps/runtime/issues/92
>



-- 
Kenneth Rohde Christiansen
Senior Engineer, WebKit, Qt, EFL
Phone  +45 4294 9458 / E-mail kenneth at webkit.org

﹆﹆﹆