Thanks for the input Steve. I think we share the same frustration
with failed attempts to bludgeon Nuance's flaky SRGS support into
submission.
Sadly, even a single CDATA is causing me problems. I was able to get
close using the compatibility hack, but was never able to assign the
correct value to it (or any value at all :-).
Unless somebody has a workable approach for getting standard grammars
to work on these platforms, I agree that it makes sense to have RDCs
optionally serve GSL. Hopefully it will just be a temporary
workaround, but it's beginning to seem unavoidable. At least it will
mean that Steve and many others can seriously consider RDCs on their
next project. Getting more eyes and hands on this should increase
the pool of expert input (I'm a newbie to the voice stuff :-).
Rahul, others, what do you think?
And does anybody have or know where we can find grammars in GSL
equivalent to the RDC basics like alpha, digits and alphanum? I've
reluctantly started looking through Nuance GSL docs, but it would be
nice if we didn't have to start from scratch.
Stu
On Jun 14, 2005, at 4:07 AM, Steve Hodson wrote:
Stu,
Sorry for being silent up until now but unfortunately due to a lack
of any real success my managers have got me working on a multi-
modal project that doesn't use RDC.
I've read through the previous messages and I can readily
appreciate Stu's frustration however all I can do is give you what
little I know.
In the link highlighted by Rahul I didn't use the compatibility
attribute - found this didn't work which at the time was nothing
new (sigh!). Instead I used instead one <tag><![CDATA[<=$a>]]></
tag>. Now you are saying that CDATA tags cause compilation
problems - me too but only when I used multiple CDATA tags. Using
one corrected the problem but I don't know why. I haven't used
either of the development environments you listed just Nuance with
sp3/4.
Might it be possible to dump the idea of using Nuance's version of
GrXML (which it seems is just a wrapper around GSL)and use instead
their GSL. I say this as struggling with Nuance GrXML may be a
fruitless enterprise if Nuance get around to supporting GrXML
properly then in theory you should be able to use the gramars
given. In the meantime writing GSL grammars may be a more
productive workaround.
Steve Hodson
VoxGen
Manor House, 21 Soho Square
London W1D 3QP
Email: [EMAIL PROTECTED]
Tel: +44 (0)870 350 2100
Fax: +44 (0)870 350 2101
Direct: +44 (0)20 7851 1025
-----Original Message-----
From: Stu Robertson [mailto:[EMAIL PROTECTED]
Sent: 14 June 2005 04:59
To: Tag Libraries Developers List
Subject: Re: RDC: Plan B for getting Nuance to work with RDCs
I would really like to find out more about Steve's success with the
grammar listed in his post. Whenever I try using CDATA in a tag, I
get a compile error in the grammar, including when I use the grammar
he mentioned worked for him. This is in the current development
environment of both community.voxeo.com and also cafe.bevocal.com.
I'm certainly at a dead end with this for now, so I guess I'll give
the group a few days to see what they can turn up. I won't redo the
grammars unless we need to, but I was really hoping to be able to get
past the compatibility issues by now. We are still unable to test
anything until 1) IBM's 6.0 Toolkit comes out for real (we can't get
the test environment/simulator to work reliably at all) or 2) we get
RDCs working on a Nuance platform, for which there are free
development environments readily available. The latter will make a
great wiki addition once we get it working.
As for deployment plans, we have both IBM's voice response/server and
a Nortel platform in house. Nortel uses Nuance grammars.
As for using different resource bundles to support different platform
grammars, will that work well if we also have separate grammars for
internationalization? I need to look at the RDC implementation a bit
more, but that's the first question that comes to mind.
G'night :-)
Stu
On Jun 13, 2005, at 4:23 PM, Rahul P Akolkar wrote:
<detour>Steve, if you're listening, can you give Stu some details
about
your Nuance setup?</detour>
Stu Robertson <[EMAIL PROTECTED]> wrote on 06/13/2005 04:23:03
PM:
Hmm. I'm starting to suspect that nuance folks mainly use GSL. I'm
still hoping for some help from the newsgroup posts I've made to
voxeo and bevocal, but I'm not counting on it.
Give that a day or two Stu. I dug up what Steve was saying here [
http://marc.theaimsgroup.com/?l=taglibs-
dev&m=111530442607124&w=2 ], and
whatever variant of the Nuance platform he was using seems to
process the
SI correctly. Are you sure the test environment you're using is upto
speed? How long will you be deploying on this setup in the future? I
suspect at some time, you'll move to an in-house arrangement? If
so, I
wouldn't spend too much time on re-doing the grammars, I'd save the
efforts for the real deal. Just my two cents worth (but, based on
personal
experience).
If we're likely to have separate grammars I'm considering just using
GSL for nuance. RDCs still abstract the app from the grammars, and
hopefully longer term we can just use SRGS for both. It's a pain
though, since we'll need to have internationalization for RDCs to be
widely adopted, and GSL would double that work.
Agreed, cross-products are no joy. Thankfully, this is search
replace type
of authoring, if it comes to that.
I agree it'd be fine to have the grammar-type/platform specified in
an init parm. The next question is where to put the grammars. What
about creating two sub dirs under .grammar, one for srgs and one for
gsl? The servlet would use the init parm to figure out which to
use. There's the issue of file name suffixes of course. If you
think this is a reasonable approach, I'll take a stab at the
necessary changes.
Sure, new directory is a good idea. But the directory structure or
suffix
has nothing to do with what grammar is picked up, it will be the
resource
bundle that dictates this. The directory or suffix will really be
just a
visual cue for developers.
For example, RDCs pick up prompts and grammars for en_US locale
from here
[
http://cvs.apache.org/viewcvs.cgi/jakarta-taglibs-sandbox/rdc/src/
org/apache/taglibs/rdc/resources/RDCBundle_en_US.properties?
view=markup
]. Lets have a locale variant called Nuance, and have that resource
bundle
point to the correct grammars (AFAIK, variants can be made up). A
brief
read on the current internationalization support in RDCs is here [
http://wiki.apache.org/jakarta-taglibs/ReusableDialogComponents/
Tutorials/I18N
].
Hopefully somebody will help us get nuance srgs to work and make the
GSL part unnecessary...though we'd still probably need the subdirs
for grammars.
Works for me, got nothing against arranging resources
systematically ;-)
-Rahul
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
######################################################################
###############
Winner - e-Government excellence 2004.
Runner up - European Information Management awards 2004:
- The Premier Project Award.
- B2C Commerce Project Award.
- CRM Project Award.
For more information visit us at www.voxgen.com
######################################################################
###############
Note:
This message is for the named person's use only. It may contain
confidential,
proprietary or legally privileged information. No confidentiality
or privilege
is waived or lost by any mistransmission. If you receive this
message in error,
please immediately delete it and all copies of it from your system,
destroy any
hard copies of it and notify the sender. You must not, directly or
indirectly,
use, disclose, distribute, print, or copy any part of this message
if you are not
the intended recipient. Vox Generation Limited and any of its
subsidiaries each
reserve the right to monitor all e-mail communications through its
networks.
Any views expressed in this message are those of the individual
sender, except where
the message states otherwise and the sender is authorised to state
them to be the
views of any such entity.
Thank You.
######################################################################
###############
######################################################################
###############
This e-mail message has been scanned for Viruses and Content and
cleared
by NetIQ MailMarshal
######################################################################
###############
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
