Re: draft charter for a W3C NFC working group

Hi Dave,

Thanks for starting this effort.

I would argue with the fact that LLCP is not widely supported. Android,
Symbian, Linux and probably Win8 do support it.

As Stephen noted, LLCP services map really well to a socket model, which is
why we implemented them as native sockets in the Linux kernel. So using Web
Sockets would make sense, in my opinion.

I think you can safely assume that you will only get NDEFs through an SNEP or
NPP LLCP service, so you can use them to read and write NDEFs from/to an NFC
device.
You have no guarantee about the kind of LLCP payload you will get from any
other LLCP service (except for the TCP and OBEX bound ones).
In other words, I think it would make sense to provide an LLCP sockets API in
order to allow non SNEP use cases to be implemented as web apps.

Cheers,
Samuel.


On Thu, May 10, 2012 at 10:49:21AM +0100, Dave Raggett wrote:
> Generally speaking it is preferable to keep working group charters
> narrowly scoped as this helps when it comes to decisions about joining
> working groups as well as with progressing specs along the standards track.
> 
> If there is a consensus that we need to address LLCP, we can easily add
> it to the charter, but we need to understand the implications. The NFC
> forum describes LLCP as:
> 
> > LLCP is a compact protocol, based on the industry standard IEEE
> > 802.2, designed to support either small applications with limited
> > data transport requirements, such as minor file transfers, or network
> > protocols, such as OBEX and TCP/IP, which in turn provide a more
> > robust service environment for applications. The NFC LLCP thus
> > delivers a solid foundation for peer-to-peer applications, enhancing
> > the basic functionality offered by ISO/IEC 18092, but without
> > impacting the interoperability of legacy NFC applications or
> > chipsets.
> 
> How would we want to expose this to Web applications? Some possibilities
> include XHR, Web Sockets, and HTML5 Message Channels. Given that LLCP is
> relatively new and not widely supported (correct me if I am wrong), it
> may be something we should leave to a later date.
> 
> Best regards,
> 
>   Dave Raggett
> 
> On 09/05/12 20:11, Stephen Tiedemann wrote:
> > Hi Dave,
> > 
> > nice work and great to see that rolling in the browser community.
> > 
> > I'm wondering though why the scope shall be limited to NDEF exchange,
> > even in P2P mode. I would expect a lot of great ideas to blossom up when
> > access to the LLCP interface becomes available, it's a much more
> > flexible way of content exchange and allows interactive communication.
> > Conceptually, an LLCP service interface matches pretty well with a
> > socket model (done in https://launchpad.net/nfcpy project, for example).
> > 
> > /Stephen
> > 
> > On Tue 08 May 2012 11:22:06 AM CEST, Dave Raggett wrote:
> >> See http://www.w3.org/2012/05/nfc-wg-charter.html
> >>
> >> Your comments are welcomed!
> 
> 
> -- 
> Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
> 

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

Received on Thursday, 10 May 2012 17:40:29 UTC