- From: KOMATSU Kensaku <kensaku.komatsu@gmail.com>
- Date: Wed, 8 Jun 2011 10:43:38 +0900
- To: Scott Wilson <scott.bradley.wilson@gmail.com>
- Cc: Dan Brickley <danbri@danbri.org>, Matt Hammond <matt.hammond@rd.bbc.co.uk>, public-web-and-tv@w3.org, Jean-Claude Dufourd <jean-claude.dufourd@telecom-paristech.fr>
Hi, I'm apologize if my senario shown below is not suitable for this thread and HNTF. I think second screen senario has to consider about not only controlling each other but also how to get real-time tv data ( such as tracking data of so ). I think people doesn't want only tv relevant widgets, they also want to google web contents which is relevant to tv program ( shopping item, travel information, nice restaurant etc). To achieve it smooth way, I think those two points shown below are worth to consider. 1. data format ( describing detailed TV program. I guess WebVTT is candidate for it, but I don't see current spec is enough or not ) 2. how to deliver ( I mean server-push technology, such as Comet or WebSocket, But scaling the service needs other communication methods, UDP based technology. Cheers, --- Kensaku KOMATSU NTT Communications. 2011/6/7 Scott Wilson <scott.bradley.wilson@gmail.com>: > > On 7 Jun 2011, at 10:08, Dan Brickley wrote: > >> On 7 June 2011 10:56, Scott Wilson <scott.bradley.wilson@gmail.com> wrote: >>> I think its important not to lose sight of the "web" part. What I'm interested in is web applications/W3C widgets being able to interact with the capabilities of the TV and the content. This means that, while UPnP is quite possibly the underlying implementation, there is still a need to expose the capabilities in a way that makes sense for web developers implementing applications in HTML5 and JavaScript rather than low-level native applications directly accessing the hardware layer, which is already possible. >> >> Yes, Web developers can be impatient creatures. I'm not convinced that >> UPnP is the right level of abstraction. >> >> That said, WebGL's recent adoption shows that Javascript allows fiddly >> low-level APIs to be encapsulated by nicer toolkits and APIs. >> >> We need imho two things: an authenticated bidirectional communication >> channel with media centres and devices playing that role. And then >> specific collections of messages and interaction patterns with those >> services. For a Web standard we can't really assume the devices are on >> the same LAN; I think this will be our biggest headaches. >> >> Re widgets, absolutely. I worked on Joost.com's original Widget >> system, which was pretty much same as W3C widgets plus some TV APIs. >> It was very promising as a platform for extendding TV, ... however >> those widgets were all on-screen rather than second screen, which >> pretty much fails to be useful -- they get in the way of the TV! If we >> can move to W3C widgets on a second screen there is absolutely huge >> potential, but to do this we need that communication channel... Once >> we've got a communication channel, the rest is relatively easy. > > Its interesting how many people I know were watching Eurovision the same way - with a laptop or mobile open with the twitter feed live. A very simple "second screen" example, but still a compelling web+tv experience. Another common example is the use of third-party TV guides to augment the EPG - which doesn't have things like social content, reviews, actor bios etc. I think these types of UCs are easy to explain in user terms, and shouldn't be that challenging to implement in a more connected way without delving into UPnP territory. > > In terms of the channel, can we make use of legacy communication channels? For example, if a widget want to trigger a change of channel on the primary screen, then there is the pairing of the device with the TV to act as a controller, and then also the user granting the widget access to the channel-switching API <feature>. The latter could be handled as per DAP and WAC[1]. > >> >> cheers, >> >> Dan >> >> >>> PS I would take a look at the CEA web binding, but like most web developers I won't spend $450 on a spec document :) >> >> Well, quite. > > [1] http://specs.wacapps.net/2.0/feb2011/core/widget-security-privacy.html > > >
Received on Wednesday, 8 June 2011 01:44:17 UTC