RE: Modification of FPR 30 - Common UA <=> SS protocol

As written, FPR30 seems to require that the UA is the middle man in all communications whether or not this is necessary.  I agree with Bjorn that we shouldn't unnecessarily box ourselves in.

Thanks


-----Original Message-----
From: Bjorn Bringert [mailto:bringert@google.com] 
Sent: Monday, December 13, 2010 3:18 AM
To: Michael Bodell
Cc: Young, Milan; public-xg-htmlspeech@w3.org
Subject: Re: Modification of FPR 30 - Common UA <=> SS protocol

I think that Milan's new wording is fine for this requirement. It
captures the essential requirement that a web app that uses the API
should work in all browsers.

I don't think that we have to mention the UA in every requirement. For
example, we have lots of requirements of the form "Web apps must be
able to do X", where it's actually the UA that carries out X. Consider
the spec for something like XmlHttpRequest
(http://www.w3.org/TR/XMLHttpRequest/). It's mostly saying things like
"[XmlHttpRequest] allows scripts to perform HTTP client
functionality". It's actually the UA that implements the API, but the
UA gets very little explicit mention in the spec.

/Bjorn

On Fri, Dec 10, 2010 at 9:38 PM, Michael Bodell <mbodell@microsoft.com> wrote:
> As I mentioned at the end of last call I'm a little suspicious of moving
> from things that describe user agent and speech server communication to
> things that describe web application and speech service communication
> (taking the UA out of the picture).  The reason this leaves me skeptical is
> we have a bunch of other requirements (security, latency, privacy, feature
> set) that imply the user agent is involved in this communication.  When we
> skip the user agent and go to a description that talks about web application
> communicating to the speech service I don't see how we maintain all of the
> other requirements.  I think fundamentally the web application makes
> requests of the user agent and the user agent also may make requests of a
> speech service (which may be local or remote, may be specified using html
> tags and attributes, or via URI references, or via browser plugins, or via
> javascript, or via new other W3C standardized APIs - the exact proposals of
> which of these work with most/all our requirements should follow only once
> we have all our requirements).
>
>
>
> From: public-xg-htmlspeech-request@w3.org
> [mailto:public-xg-htmlspeech-request@w3.org] On Behalf Of Young, Milan
> Sent: Friday, December 10, 2010 12:43 PM
> To: public-xg-htmlspeech@w3.org
> Subject: Modification of FPR 30 - Common UA <=> SS protocol
>
>
>
> The existing requirement reads: "The communication between the user agent
> and the speech server must require a mandatory-to-support lowest common
> denominator such as HTTP 1.1, TBD."
>
>
>
> I would like to suggest the following modification:
>
>
>
> Summary- Web applications must be allowed at least one form of communication
> with a particular speech service that is supported in all UAs.
>
>
>
> Description - It may be difficult to agree on a common protocol that works
> for all requirements in the cross product of web applications, UAs, and SSs
> in the first version of the recommendation.  Since most "real world"
> applications require a tight coupling of the speech service and application,
> this aspect of the standardization has priority.
>
>



-- 
Bjorn Bringert
Google UK Limited, Registered Office: Belgrave House, 76 Buckingham
Palace Road, London, SW1W 9TQ
Registered in England Number: 3977902

Received on Monday, 13 December 2010 20:22:03 UTC