W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2006

RE: Field without a prompt

From: Petrova Anastasia <Anastasia.Petrova@bercut.com>
Date: Thu, 19 Oct 2006 12:20:12 +0400
Message-ID: <CD481F25862BDF4EB3014619BACB38D803627067@master.office.bercut.ru>
To: "Petr Kuba" <kuba@optimsys.cz>, <www-voice@w3.org>

I have a library that validates the scripts. It just goes through tags
and says whether everithing in the script is valid or not. Assuming that
the field is the only one in the form it will be visited again and again
(it is possible that there are no factors that prevent visiting this
field again) till the user says something matching the grammar. And this
may cause a problem since it may become an infinite loop. But it seems
that it is the correct behavior according to the spec. Thank you very
much for the answer.

-----Original Message-----
From: Petr Kuba [mailto:kuba@optimsys.cz] 
Sent: Thursday, October 19, 2006 12:01
To: www-voice@w3.org
Cc: Petrova Anastasia
Subject: Re: Field without a prompt

Hi Anastasia,

The spec seems to be very clear about behaviour in the situation you
described. As Teemu Tingander already explained, when the interpreter
should reprompt and there are no prompts in the respective <field>,
nothing
will be said to the user when interpreting the prompt.

If there is no but the default nomatch event handler present and the
user
says something unrecognizable, platform default prompt informing about
the
nomatch situation is spoken to the user and the field is visited again
(unless other factors will prevent it - see FIA in the spec), which is
an
acceptable behaviour.

If there is no but the default noinput event handler present and the
user
is silent until the noinput timeout elapses, no prompt is spoken at all
and
the field is visited again (unless other factors will prevent it), which
is
kind of pathological behaviour since the user has no feedback. Anyway,
it
is the correct behaviour according to the spec.

I don't understand why you should translate such a script for a platform
that will perform it. Such a script is a correct one, though not very
user
friendly. Could you please be more specific about your needs and the
platform
you use?


Petr

-- 
   Petr Kuba, Project Manager
   OptimSys, s.r.o
   kuba@optimsys.cz
   Tel: +420 541 143 065
   Fax: +420 585 750 429
   http://www.optimsys.cz



Petrova Anastasia wrote:
>
> Actually, I guess I need to provide better explanation of the 
> situation. It's not about writing a script where field doesn't have a 
> prompt. It's about translating such a script for the platform that 
> performs it. I actually don't know why someone may need field without 
> a prompt, but as you confirmed, it is possible. And it's also possible

> that no <catch> elements are specified within the field. So, in 
> accordance with table in section 5.2.5, if nomatch or noinput event 
> occures I need to reprompt. And the most interesting question is what 
> should I do in this situation, when I have a script with field that 
> has no prompt and no event handlers, and I need to reprompt?
>
>
------------------------------------------------------------------------
>
> *From:* Teemu.Tingander@tecnomen.com 
> [mailto:Teemu.Tingander@tecnomen.com]
> *Sent:* Wednesday, October 18, 2006 12:00
> *To:* Petrova Anastasia; www-voice@w3.org
> *Subject:* RE: Field without a prompt
>
> Hi !
>
> Without not going into the reasons reasons to do field item without 
> prompt;, Yes. Field item can exist without any prompt. In fact the 
> prompts before the collection to that fields starts can be queued from

> other elements processed during FIA collect phase. But if you indeed 
> have a field item without any prompts and reprompt request occurs 
> nothing is putted to prompt queue and thus nothing is prompted. In 
> this case you should use event counts to define amount of events 
> like"nomatch" can occur in that field. In my example this count is 3.
>
> <nomatch count="3">
>
> <-- executed when event nomatch is thrown 3rd time while the 
> collecting this field -->
>
> <prompt> Sorry I could not really under stand you, bye </prompt>
>
> <-- lets disconnect the phone call-->
>
> <disconnect/>
>
> </nomatch>
>
> You should probably do something more clever in here. Studying FIA 
> helps a lot to understand the VoiceXML process / collect loop that 
> plays a big role in this question too. If form has more field items 
> that you want to collect you may use cond or filling this particulars 
> field item to make FIA to skip its collect.
>
> The interesting part in here is the "timeout", but that's another 
> story then.
>
> - Teemu
>
> (oh I see that you got some answer already but..)
>
>
------------------------------------------------------------------------
>
> *From:* www-voice-request@w3.org [mailto:www-voice-request@w3.org] *On

> Behalf Of *Petrova Anastasia
> *Sent:* 17. lokakuuta 2006 11:22
> *To:* www-voice@w3.org
> *Subject:* Field without a prompt
>
> Hello. My name is Anastasia and I have a question about the
specification.
>
> In 2.4.1 there's a sentense: Each field can have one or more prompts.
>
> Can you clear to me, whether it is possible that field have no prompts

> at all?
>
> If this is really possible, what should I do when nomatch or noinput 
> event occures and I have to reprompt?
>
> Is it an infinite loop till user says something matching a grammar, or

> I should throw an error?
>
> Thank you in advance.
>
> //Best regards,//
>
> //Anasatsia Petrova//
>
> //Software Analyst//
>
>
> <http://www.bercut.com/>
>
> // //
>
> /liter '3', 22 Moiseenko street,
> St. Petersburg, Russia, 191144
> Phone/fax: +7 812 327 32 33,
> E-mail: Anastasia.Petrova@bercut.com
> Web-site: //http://www.bercut.com/
>
Received on Thursday, 19 October 2006 08:20:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:49:04 GMT