W3C home > Mailing lists > Public > public-web-intents@w3.org > May 2012

Re: Web Intents for local network services (DAP Action-510)

From: Dave Raggett <dsr@w3.org>
Date: Fri, 25 May 2012 12:44:55 +0100
Message-ID: <4FBF70B7.8040200@w3.org>
To: "Nilsson, Claes1" <Claes1.Nilsson@sonymobile.com>
CC: "Cathy.Chan@nokia.com" <Cathy.Chan@nokia.com>, "public-web-intents@w3.org" <public-web-intents@w3.org>
On 25/05/12 12:04, Nilsson, Claes1 wrote:
> Hi again,
> It seems as the requirement in the UPnP Device Architecture 1.1 specification that each discovery message MUST fit entirely in a single UDP packet, that might be down to 512 bytes, prohibits us to define long headers that perform full Web Intents Service registrations. So currently I see no other option than relying on the Device Description document, as proposed in Shenzhen, for Service registration. 
> In addition to this we could just add one extra header "action.webintents.org:...." that helps UA's Web Intents implementation to filter SSDP responses. It would be good to have a " type.webintents.org:.." header as well but then we risk to exceed the limit of 512 bytes if several types are supported.
> What do you say. Any views?

First: It would be helpful to know which devices supported web intents
so that you only need to retrieve the service descriptions from those
devices that do support web intents.

Second: An alternative to extending the UPnP service description format
would be to define a new format that is returned separately and used to
describe the parameters for web intents. It would be convenient is this
used the JSON format. The browser wouldn't need to retrieve the XML
service description file until that becomes necessary.

Combining these two, we arrive at a single new header for the SSDP
response/advertisement that references the location for the web intents
description. If the heading is missing, the device isn't web intents aware.

Have you done any data collection on the size of SSDP responses and
unsolicited advertisements for current devices? If it turns out that
these are typically small, then 512 bytes might give us sufficient room
to further ideas.

Best regards,
Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
Received on Friday, 25 May 2012 11:46:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:14:46 UTC