| Re: [vtp-dev] Upcoming Release of 2.0 |
Trip -
The Tellme Launcher is my baby. I created it for VTP 1.0. The version that is available in the last daily build of 1.0 unfortunately is quite broken when used with current versions of Skype. I have checked in a fix already, and the binary I have built myself at home works fine. So, to keep the Tellme Launcher functionality, at one level, just requires a rebuild of the plugin. The plugin was implemented using Java 1.4, and with more recent Eclipse cores, it would be better raised to 1.5 level. I have a version running on Eclipse Callisto (3.2) just fine.
As for whether the Tellme Studio service will continue now that Tellme is part of Microsoft -- I have asked that question of contacts I still have there, and the answer is "no change has been contemplated". Who knows how long that will continue? I don't see it changing very soon, though. Tellme devoted almost zero resources to Studio over the past 3-4 years, and it just keeps working. The last changes made to Studio were made by me during the last few months before I retired. Tellme continues to run its VXML platform, and I see no reason why they would change that, and so long as the platform runs, Studio kind of comes along free.
So, I think continuing to include the launcher functionality is viable. I will be happy to continue working with it, if you like. Have you got any notions for enhancements/features that the launcher should have? There are two things that I can think of: 1) Getting Linux and Mac OS versions of the launcher created. Right now it only runs on Windows. 2) During the past year, Skype has itself put some effort into creating its own blessed interface code to connect Java applications to the Skype client. It might be well worth exploring using those facilities instead of the homegrown ones I did. I can see potential difficulties with IP issues with that code inside the Eclipse Foundation. I am not sure that Skype was real "clean" in its licensing of that code.
Your idea of bucking the VXML editor up to the WTP sounds like a very good thought, and will take those editors out of the VTP resources' hair, and leave the functionality to the end user. Do you think you can do that successfully?
-- Mike
----- Original Message -----
From: Trip Gilman <mailto:trip@xxxxxxxxxxxxxxx>
To: Vtp-Dev <mailto:vtp-dev@xxxxxxxxxxx>
Sent: Friday, September 14, 2007 3:33 PM
Subject: Re: [vtp-dev] Upcoming Release of 2.0
Mike -
I’ve been thinking about those very things for some time now. I’ll try to break down my thoughts on the features that will likely survive. Of the features provided in the VTP 1.0 release, here are the ones I believe will/should survive:
- Grammar Authoring – Constructing grammar files is an integral part of voice application development. I don’t believe our project could be considered complete without it. To that end, a GRXML editor will be included in the 2.0 release in one form or another. However, I have concerns with continuing to have a dependency on the WTP. I see two options here, first we recreate the editors with like functionality using some other xml editor core or try to promote the xml editor core higher up the stack than the WTP (many other projects have their own cores for this very reason). Our second option is to continue with our requirement on the WTP (only for the xml editors, which is overkill I think) but we will need to perform some maintenance to provide compatibility with the latest edition of the WTP.
- TellMe Launcher – I would love to have this functionality survive our transition into 2.0. To that end, does anyone have the inside scoop on whether this service will continue to be available under Microsoft’s flag? Also, the OpenMethods boys don’t have the expertise to truly maintain this section of the code base. Anyone want to volunteer as the defacto maintainer of this?
As for the rest of the VTP 1.0 feature set, the 2.0 release offers at least comparable, if not enhanced, functionality. It is true that the 2.0 release will supplant the current code base in many areas. I believe this will significantly lower the bar of entry for users who are not of the developer persuasion, and provide capabilities that are just not maintainable when the ultimate output of the tooling is raw VXML. Developing multi-voice or multi-language applications can be very time consuming. If these details can be managed by our tool transparently to the user while both developing the application and during its eventual deployment, I’m all for it.
My ultimate hope for this project is that it will provide all the features necessary to develop an application from its inception to its grave. Many development tools take a “get close enough, and let the programmer finish it” approach to application design. We have the expertise and the capabilities as a group to not just get close enough, but take that next step and create a truly user friendly development environment for authoring voice applications.
I’ve also been considering what to do with the VXML editors that are currently in the VTP 1.0. As a developer, I just hate to throw perfectly good code away. As you said in your email, there are still many projects and/or cases where a programmer wants to be “close to the metal”. What do you think about a cross project contribution? I suggest offering the VXML editors up to the WTP as an additional language specific editor set. In this way, the developers who are just creating a web application that happens to use VXML instead of HTML still have that capability. It would also be more in line with the path being taken by the WTP development model.
Thoughts?
Trip Gilman
**thread truncated for brevity**
On 9/12/07 12:27 AM, "Mike Greenawalt" <mike@xxxxxxxxxxxxxx> wrote:
It is great to begin seeing the eclipse.org/vtp website change to reflect the advent of a VTP release based on openVXML. I see a lot of work has been done.
As a contributor to VTP 1.0, I have been very interested in understanding how the openVXML-based VTP would relate to the original VTP. It's beginning to be clear that this version will completely supplant the previous version. Is that correct? Will there remain any overt support for creating and editing VXML and grammar files? Application launching?
I have used VTP 1.0 for several small projects, and find it tremendously useful still. Is there any sense in which its capabilities will survive? There are many, many voice application projects that can still be well-serviced by the capabilities that were in VTP 1.0.
-- Mike
_______________________________________________
vtp-dev mailing list
vtp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/vtp-dev
_______________________________________________
vtp-dev mailing list
vtp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/vtp-dev