Fw: action eric, look at Semantic Interpretations for Speech Recognition

Hi, all

Here forwarded is my message to EricP, concering
the Semantic Interpretations for Speech Recognition document.
(This is done based on Eric's general permission to forward his message)

After receiving Steve's message
http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0302.html
I think my analysis may contribute to the discussion,
for the role of the ECMAScript in my understanding is slightly different.

Best,
Yoshio

----- Original Message ----- 
From: "Yoshio FUKUSHIGE" <fuku@w3.org>
To: "Eric Prud'hommeaux" <eric@w3.org>
Cc: <fuku@w3.org>
Sent: Tuesday, August 24, 2004 7:38 PM
Subject: Re: action eric, look at Semantic Interpretations for Speech
Recognition


> Hello Eric,
>
> > Yoshio, it looked to my like folks were using a grammar language to
> > have the speach recognizer construct either Ecmascript or XML
> > structures. At first glance, does that seem right to you?
>
> Well, there seems to be two interpretation to "that" in your question.
>
> (1) that = using a grammar language ...
> (2) that = it looked to ...
>
> However, in both case, my answer is "Yes." ; )
>
> # It is very common to include some codes (scripts, directions, whatever
you
> want to call them)
> # to produce a construct of a sentence (utterance, in this case).
>
> However, the result in XML form seems to me to be claimed as a
> derivation from the Ecmascript one.
>
> And they allow the (abstract) grammar lanugage expressed both in ABNF
> and XML.
>
> > My action was to determine the relevence to the DAWG. Apart from the
> > fact that everything *could* be expressed in RDF, this spec doesn't
> > seem to overlap. The rules seem very high (specialized-application)
> > level and the parsed semantics are identified only by Ecmascript
> > variable names rather than URIs.
>
> Well, IMHO, I don't understand what you mean, sorry.
>
> What the spec may have anything to with DAWG is, in my understanding,
> their position of having an abstract grammar language
> and allowing both ABNF and XML as concrete languages.
>
> However, while Ecmascript is used as the common result format,
> no converter  from the ABNF grammar to XML grammar (and vice versa)
> is presented there.
>
> So, this spec just provides us an example of  having a common
> abstract language (and a common format for the result ),
> while having two (distinct) format as the concrete ones.
>
> > But I'm not an NLP guy. Maybe I'm just crazy.
> > Could you take a quick glance and tell me if I am?
>
> If you are crazy? I'm not sure.
> Do you think a crazy can tell if other person is a crazy? : )
>
> This is the limit of my understanding for now.
>
> Best,
> Yoshio
>
>

Received on Tuesday, 24 August 2004 11:03:48 UTC