- From: Giuseppe Pascale <giuseppep@opera.com>
- Date: Tue, 26 Apr 2011 15:04:29 +0200
- To: "Bob Lund" <B.Lund@cablelabs.com>, "Jean-Claude Dufourd" <jean-claude.dufourd@telecom-paristech.fr>
- Cc: "public-web-and-tv@w3.org" <public-web-and-tv@w3.org>
** added ISSUE-7 to the subject for tracking purposes *** On Wed, 13 Apr 2011 09:54:23 +0200, Jean-Claude Dufourd <jean-claude.dufourd@telecom-paristech.fr> wrote: > On 12/4/11 16:51 , Bob Lund wrote: >> >>> -----Original Message----- >>> From: public-web-and-tv-request@w3.org [mailto:public-web-and-tv- >>> request@w3.org] On Behalf Of Jean-Claude Dufourd >>> Sent: Tuesday, April 12, 2011 6:12 AM >>> To: public-web-and-tv@w3.org >>> Subject: Re: [HOME_NETWORK_TF] Some use cases and requirements for >>> broadcast TV applications >>> >>> On 11/4/11 21:52 , Bob Lund wrote: >>>> UC4 seems like a web server function: UI from server X moves from >>> client A to client B. This could be done by a web server with HTML5 >>> within a home network or across the internet. Nothing new is required, >>> I >>> think. UC5 has the element of device A discovering a service on device >>> B >>> and device A creating UI presented on device B. >>>> Bob >>> JCD: UC4 is about moving the interface (HTML page plus all its >>> ressources) *plus* the current state of the interface, such as which >>> choices are already made. >>> So the question is, where does the state reside. >>> >>> Option 1: You have an intelligent server on the Internet, and the web >>> page on the terminal is progressively updated using AJAX. The state >>> information is on the server, and the terminal is just a mirror of this >>> information. It is feasible indeed to connect from another terminal, >>> identify as the same user, and get the intelligent server to switch to >>> using the new terminal for the already started session. >>> >>> Option 2: The intelligence is in the web page (widget-style). The >>> server >>> is plain HTTP. The state information is in the page. The you have no >>> way >>> to transfer the state information. You can restart the interface >>> elsewhere, but it starts "at the beginning". >>> This could actually be an interesting requirement to contribute to the WG working on Widgets, i.e. being able to "move" a widgets among devices (including the state of the widget) >>> >>> I claim that option 2 is useful and needed, for example to satisfy >>> privacy concerns, hence my proposal for UC4. >>> Best regards >>> JC >> I agree these are the options. I think option 1 can be implemented >> exclusively through JavaScript so would lend itself to deployment to >> any existing browser. Agree, so this option can be ignored for our purpose. > For option 2, are you proposing a HN service on >> the target client that the source client invokes and transfers its UI >> state to? If so, I would agree this fits within the HNTF scope. > JCD: Yes, I am proposing a mechanism to transfer UI state. You sketch a > push mechanism, whereas we have currently implemented a pull mechanism. > The sender advertizes the URL of the UI state, and the receiver pulls > the UI state off the URL. > Actually, the full implemented mechanism is: > 1- all possible target client advertizes themselves as providers of a > service called "home network user agent" > 2- the source client discovers the target client > 3- the source client binds to the target client then sends a message > with the URL of the content and the URL of the UI state > 4- the target client fetches the UI content then the UI state > Because we are using UPnP, this pull solution is easier. Some comment on the current text: ** maybe would be better to change "document" with something more generic In general I see 2 things we want to be able to move: services and content. The "document" case is probably a but ambiguous since it could be regarded both as a service and as content. In your case would be probably better to use service though, so saying somethin like "A service moving across devices", "how do we move a service and its current "state" to another device" etc. So a document would be a specific case for a more general case of service migration. This also imply that you need a mechanism for devices to negotiate a supported "service" type What do you think? @Bob, Would this address also your comment? Note that this would require a clear definition of "service" that should include things like, for example, a Remote UI. ** please try to fill in the "Need/justification" section, that really important to progress the discussion and have final approval of the use case. /g > Best regards > JC >> Regards, >> Bob >>> -- >>> >>> JC Dufourd >>> Directeur d'Etudes/Professor >>> Groupe Multimedia/Multimedia Group >>> Traitement du Signal et Images/Signal and Image Processing Telecom >>> ParisTech, 37-39 rue Dareau, 75014 Paris, France >>> Tel: +33145817733 - Mob: +33677843843 - Fax: +33145817144 >>> > > -- Giuseppe Pascale TV & Connected Devices Opera Software - Sweden
Received on Tuesday, 26 April 2011 13:05:32 UTC