Martin,

Glad to get your views. I feel the RDC taglib meshes in closely with the 
following:

1. JSTL:  RDC tags are JSP 2.0 tags that are currently, and we expect this 
trend to continue, commonly implemented using JSTL rather than Java.  The 
reasons include the common ones for adopting JSTL, ease of use for 
prototyping, ability for tools to generate code easily, common programming 
model for tag use and tag definition.

2. Struts: The RDC framework is geared to making the JSP author's job 
easier for building voice applications. We include several helper tags in 
addition to integrate with Struts. There's a <struts-submit> tag which 
functions like form submits in HTML that passes data collected in the JSP 
to a struts form bean for validation and action processing.

3. XTags and I18N taglibs: RDC tags need to process XML configuration 
files giving prompts, error strings, help messages etc. and to apply those 
to the tag being used. XPath and to some degree XSLT are being used for 
this and there is good opportunity to look at XTags as well. Clearly, 
voice apps must be internationalized atleast as well as visual ones.

4. Request, Response, Session and Application taglibs: A voice app's 
interaction in these contexts is very similar to conventional web apps and 
hence, there will be the same level of synergy.

The RDC taglib builds on existing Taglibs efforts, and in addition brings 
some interesting new function to the web. The RDC tag library is designed 
to play a synergistic role when creating voice views of web applications, 
like struts does in the conventional web apps. Since a component is 
reusable, can collect arbitrarily complex information and encapsulates 
behavior (validation, canonicalization etc.), an individual component can 
be looked at as a "tile".

One main difference between voice JSPs and visual ones is the need to 
control user focus during voice applications. In a visual app it's great 
to have the entire JSP render and let the user decide which form fields 
he's going to interact with. In a voice application, the JSP author 
typically will want to guide the user's focus by controlling which part of 
the page renders on each server turn as a function of where the user is in 
the interaction -- sort of a within-page form of dialog management.  The 
<group> tag in the RDC framework works like a voice-equivalent of a 
"layout" manager in a conventional GUI toolkit -- and gives focus (i.e. 
permission to render) to successive child tags as a function of how much 
work the user has completed so far.

I understand that some of us might not be familiar with building voice 
applications, but when I look back, I can remember of a time when I wasnât 
in my "comfort zone" writing a conventional web app. I tried, I often 
failed, I learned, Iâm still learning, but Iâm glad I jumped in ;-)  I 
invite you to do the same with RDC tag library, and the next time Iâm sure 
youâll say you know more than diddly about voice applications ;-) That is 
exactly the advantage of the RDC tag library, bringing home to the 
conventional web app developer the ability to author voice applications 
while reusing most of the web app development knowledge related to user 
task flows, business process flows, enterprise beans, and back-ends/data 
sources. By starting to bring together the web developer community and the 
voice developer community through the adoption of a common, and existing, 
programming model for both -- namely JSPs and tag libraries, we hope we 
can add value to both communities.

-Rahul





Martin Cooper <[EMAIL PROTECTED]> 
09/08/2004 09:01 PM
Please respond to
"Tag Libraries Developers List"


To
Tag Libraries Developers List <[EMAIL PROTECTED]>
cc

Subject
Re: PROPOSAL New RDC tag library






On Wed, 08 Sep 2004 19:01:26 -0400, Sam Ruby <[EMAIL PROTECTED]> wrote:
> Per the New Project Submission guidelines[1], the next stage after
> Proposal is Sponsorship.  I'm willing to sponsor this project through
> the sandbox.  What that means is that I'm now looking into obtaining
> software grants and/or CLAs, and filling out the incubator
> ip-clearance-template information, and committing the code into the 
sandbox.
> 
> If anybody has any concerns or questions, please let me know.

My primary concern is that there appears to be zero interest in this
from the existing Jakarta Taglibs community. If none of the proposed
committers for this taglib are existing Taglibs committers, then I
foresee RDC being its own little world, with little interaction with
the rest of Taglibs. I see Taglibs already as a (sub-)project that
needs more synergy between the components, or at least within the
community, and it seems to me that this (RDC) might lead us in the
opposite direction, instead of helping to unify/unite us.

(I understand that the project won't necessarily land here after a
successful incubation, but I'm not sure I see any other candidate
locations within the ASF. ;)

Frankly, I'm also not all that clear on how similar or different a
voice-based tag library is from those that are used to build
"conventional" web apps. On the surface, it sounds like a whole
different ball of wax, but then I know diddly about building voice
applications. ;-)

My 2 cents...

--
Martin Cooper


> 
> - Sam Ruby
> 
> [1] http://jakarta.apache.org/taglibs/newprojectsubmission.html
> 
> ---------------------------------------------------------------------
> 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]



Reply via email to