W3C home > Mailing lists > Public > public-xg-htmlspeech@w3.org > October 2011

Re: Speech-enabled form filling example

From: Glen Shires <gshires@google.com>
Date: Thu, 27 Oct 2011 08:57:30 -0700
Message-ID: <CAEE5bcgBeioxY6berA48-FE9WVpSZeF40NcbqMLyHEWAkMHJ=A@mail.gmail.com>
To: Charles Hemphill <charles@everspeech.com>
Cc: public-xg-htmlspeech@w3.org
Nice use case.

A UA could implement a Voice IME using the type of speech navigation
depicted in the video. That is: user can speak a field label to focus on the
field, and speak a field value to fill-in the focused field.  (But this
behavior would be UA-dependent, and thus I believe beyond the scope of this
group's proposal.)

What is in scope of this proposal is how a <evsp:grammar> or <reco> tag can
be used to supply grammars and other information to the recognizer.  In this
respect, I believe this is remarkably similar to my <reco> proposal.  For
example, your snippet of markup could also be written as:

            <label for="KIDNEY_FAT_id">Kidney Fat</label>
            <reco
addGrammar="builtin:input?type=range&min=0&max=1&step=0.01">
              <input type="text" size="6" maxlength="4"
                     name="KIDNEY_FAT" id="KIDNEY_FAT_id" value=""
                     onchange="validateField(this, true);" />
            </reco>

Or better yet, using HTML5 validation, no <reco> or <evsp:grammar> tag would
be required as the grammar is implicit.

            <label for="KIDNEY_FAT_id">Kidney Fat</label>
              <input type="number" min="0" max="1" step="0.01"
                     name="KIDNEY_FAT" id="KIDNEY_FAT_id" value="" />
Received on Thursday, 27 October 2011 15:58:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 27 October 2011 15:58:34 GMT