W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2004

Termchar behaviour

From: <Teemu.Tingander@tecnomen.com>
Date: Thu, 18 Nov 2004 13:53:20 +0200
Message-ID: <BC70F0884912B54E9F65043933CFF575011577B7@aunty.tecnomen.fi>
To: www-voice@w3.org

Dear VBWG.

(And Hi to everyone..)

I have a simple questions about 'termchar' and I know that there is some old
discussions available about this issue in archives. But I hope that I can
get an all-clear-up answer for this (especially from you guys in VBWG, How
did you planned this to be not how it is implemented in currently available
interpreters?).

"Is termchar a input or is it just a way to terminate timeouts and speed up
processing ?"

My opinion is "terminate timeouts and speed up processing".

I was thinking that it should go this way feel free to correct me if I'm
wrong: I Think this is quite that same thing that Deusdit Correa Cornejo
pointed out in
http://lists.w3.org/Archives/Public/www-voice/2004JulSep/0040.html

So in grammar that collects (digits);
    - If no characters are inputted and grammar is not ready to terminate.
Pressing termchar would result to noinput (this makes analogue of staying
quiet in speech)
    - If no characters are inputted and grammar is ready to terminate (like
digits?minlength=0). Pressing termchar would result to fielditem to be
filled with empty string fieldItem = { '' } ; This makes it little bit
tricky on speech side but to make analogue (and depending how grammar is
made ) this kind of grammar newer throws
noinput. But this is a choice that application developer can make.
    - If characters are inputted and grammar is not ready to terminate
nomatch is thrown, this Is how app developer makes difference if user
entered invalid input or noinput at all..
    - Then As in appendix..

There is always possibility to set the termchar to something like DTMF 'A'
which is really hard to be pressed on normal phones or to empty which then
terminates termtimout but this the removes some other possibilities
   
If application developer really wants to know if collection was ended by
pressing termchar he is free to use grammars containing termchar. In
application flow on VXML side this does not seem to make any difference..

In VXML2.1 there should be analogue shadow variable ($.termchar) in fields
whose collection termination timeout was ended with '#';

So in record ;(assuming that silence detection works)
    - if no input (passed silence detection) is collected and termchar
(dtmfterm) is pressed noinput should be thrown ,but this collides the  VBWG
statement in
http://lists.w3.org/Archives/Public/www-voice/2003OctDec/0024.html. And then
the statement makes this analogue with complex grammar that has COLLECTED
SOMETHING ( and this sis where I think the difference is ) form user but
does not assign the form input item variables My opinion is that the
termchar and matching grammars returning empty are different thing. As well
as that termchar is not input in input means..

I hope that I got all my thoughts here but I doubt it.. Hopefully this still
does some sense all to you and I'll get a answer of some kind.

Best regards,
- Teemu
Received on Thursday, 18 November 2004 13:01:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:49:00 GMT