W3C home > Mailing lists > Public > public-web-and-tv@w3.org > November 2011

Re: [hntf] Summary of the joint session on Device discovery

From: Vickers, Mark <Mark_Vickers@cable.comcast.com>
Date: Mon, 14 Nov 2011 18:23:45 +0000
To: Giuseppe Pascale <giuseppep@opera.com>, Russell Berkoff <r.berkoff@sisa.samsung.com>
CC: "public-web-and-tv@w3.org WG" <public-web-and-tv@w3.org>, Robin Berjon <robin@berjon.com>
Message-ID: <27688840-A022-4408-B0FD-4EA312A62DAB@cable.comcast.com>
I'm very supportive of the joint discussion between the three efforts: CableLabs/Opera, Webinos and Web Intents for several reasons:
- While the three APIs are not at all the same, there are indeed overlaps in functionality between the efforts
- Even if these don't end up the same API, it makes sense for a consistent web API to have them all designed in a similar manner.
- Browser vendors and other W3C members have voiced a number of concerns about home network discovery APIs, including security, fingerprinting, and web app model. We need to go into this process with an open ear to these concerns and an open mind to new approaches. This is the best way to get these APIs adopted in products.

I expect we'll have a similar set of issues any time IG requirements transition to a WG API effort.

Thanks,
mav


On Nov 14, 2011, at 5:23 AM, Giuseppe Pascale wrote:

> Hello Russel,
> I'm sorry but I fail to see why the current discussion/outcome is against the agreement reached in the HNTF group.
> 
> The agreement of the group so far was to bring our use cases for discussion to the DAP group and discuss with them the best way forward.
> Such discussion happened in the joint F2F at TPAC. During this F2F all the people in the room agreed that there could be overlaps between our use cases and web intents,
> so it was decided to investigate further such potential overlaps and decide later what to do.
> 
> As said such overlaps are only about exposing services to the application (since this is what web intents is providing).
> The actual communication will be a topic for a different work, but also in this case since the DAP group is addressing similar challenges for other APIs, there is benefit in keeping it in DAP to aim for a similar approach and also to leverage the expertise in that group.
> 
> So for the time behind it seems to me that makes a lot of sense to keep the discussion in one place. If the group later decides to split it, that will be a decision of the group.
> The HNTF was chartered to identify use cases and bring them to a WG. Now the follow-up is in the WG hands, so if you have technical concerns feel free to join DAP and make your comments there.
> 
> 
> Best Regards,
> /g
> 
> 
> On Mon, 14 Nov 2011 10:26:25 +0100, Russell Berkoff <r.berkoff@sisa.samsung.com> wrote:
> 
>> Hello Giuseppe,
>> 
>> I have some concerns about the sudden appearance of a nexus between Web Intents and Home Network requirements in developed in the HNTF.
>> 
>> I donít recall Web Intents topic ever being discussed in HNTF and would have some significant concerns with merging the Web Intents and Home Network scopes.
>> 
>> If Web Intents aspires to be compatible at some level with Home Networks thatís fine. If there is a suggestion that Web Intents is a primary (only) discovery mechanism for home-networks thatís not so fine.
>> 
>> Let me start by saying that Web Intents service-centric architecture is contrary to Home Networks device-centric architecture. This will potentially limit the types of HN devices which could be offered.
>> 
>> I think HNTF may want to continue to meet until there is a consensus on the disposition of the TF's work!
>> 
>> Regards,
>> Russell Berkoff
>> 
>>> -----Original Message-----
>>> From: Giuseppe Pascale [mailto:giuseppep@opera.com]
>>> Sent: Monday, November 14, 2011 12:24 AM
>>> To: public-web-and-tv@w3.org
>>> Subject: [hntf] Summary of the joint session on Device discovery
>>> 
>>> Hi all,
>>> sorry for the late report.
>>> 
>>> This is a summary of the joint session on Device discovery between Web
>>> and TV IG and DAP WG during TPAC 2 weeks ago.
>>> If something is not clear, feel free to ask.
>>> 
>>> - I set the context of the home network task force, highlighting
>>> identified gaps that the DAP WG now needs to address:
>>>  * discovery/advertisement to allow Web applications to be aware of
>>> services available on the home network, and/or to advertise themselves
>>> on the home network.
>>>  * communication: need to relax same origin restrictions (when CORS is
>>> not
>>> supported)
>>> See presentation at:
>>>  http://people.opera.com/giuseppep/hntf/
>>> 
>>> - Clarke presented a demo of the CableLabs/Opera proposal
>>>  http://www.w3.org/mid/CAE1B1EF.14108%25c.stevens@cablelabs.com
>>> 
>>> - Dave Raggett presented a demo of the Webinos proposal
>>>  http://www.w3.org/mid/4EBCF297.8070203@w3.org
>>> 
>>> - The discussion focused on security concerns triggered by exposing the
>>> local network to Web applications, and on possible mechanisms to reduce
>>> the surface of attack. Some security policy needs to be applied between
>>> discovery requests and responses before attempting device pairing and
>>> control. Exposing service details such as its address on the local
>>> network or its exact identity may not be needed: the UA could perhaps
>>> rather expose an opaque object to the JavaScript to communicate with
>>> services identified by the type of service they fulfill. Concerns about
>>> the idea of allowing application to coomunicate with UPnP devices
>>> diretly since these devices may have not been designed for this
>>> scenario so probably not enough tested for this scenario.
>>> 
>>> - Web Intents proposal was presented
>>>  www.weintents.org
>>> 
>>> - The Web Intents proposal seems to be going in a similar direction,
>>> exposing services to Web applications through a broker.
>>> 
>>> - No decision was taken to adopt/reject any of the proposals. DAP will
>>> work on Web Intents per its charter. It is unclear at this stage
>>> whether Web Intents solves discovery requirements identified in the
>>> Home Networking Task Force, in particular in terms of supporting the
>>> protocol supported by the device that exposes the service. The two
>>> efforts may not result in a single API in the end.
>>> 
>>> - DAP will work on Web Intents in a dedicated task force and specific
>>> mailing-list and, pending WebApps gets re-chartered, as a joint
>>> deliverable between WebApps and DAP. Discussions on device discovery
>>> should take place in that task force as well, unless it appears clear
>>> that two APIs need to be defined. On-going discussions on DAP main
>>> mailing-list might conclude one way or the other, or leave it up to the
>>> task force to solve the issue:
>>>  http://lists.w3.org/Archives/Public/public-device-
>>> apis/2011Nov/0054.html
>>> 
>>> 
>>> CONCLUSION:
>>> Participants of the Home Network TF are encouraged to join DAP, where
>>> further discussions will take place. We are not going to make any work
>>> in this IG unless required at some point in future.
>>> Thanks to all TF members for their contribution and looking forward to
>>> your contribution in DAP.
>>> 
>>> 
>>> --
>>> Giuseppe Pascale
>>> TV & Connected Devices
>>> Opera Software
>> 
> 
> 
> -- 
> Giuseppe Pascale
> TV & Connected Devices
> Opera Software
> 
Received on Monday, 14 November 2011 18:24:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 14 November 2011 18:24:55 GMT