Home » Archived » WTP - PAVE » Comments on Project Proposal
|
| Re: Comments on Project Proposal [message #468615 is a reply to message #468613] |
Wed, 20 May 2009 07:18   |
 Dimitar Giormov
Messages: 13
Registered: July 2009 |
Junior Member |
|
|
Hi Konstantin,
First we absolutely agree that the framework should in the end go in the
platform. But we also think that the WTP Data Model framework should go
there. The PAVE framework depends only on the data model framework from
WTP, so we intend to create a separate downloadable with DMF as a first
step.
And if the PAVE framework gets enough attention we can try to push it in
the platform.
I will try to explain our vision for the framework, thus I think answering
most of your questions.
The idea behind the framework is actually easy sequencing of operations
existing and non-existing ones. We want to build a data model with all of
the data needed for all operations and then execute them in a single pass.
Also UI should be reusable, one and the same page could be used for
gathering data for creating a java class for instance and in the same time
to be reused in one of the steps of a certain pattern. Reusability is one
of our top priorities.
We want these operations to be able to be executed in headless mode, so
there can be multiple actions for executing one and the same pattern (drag
and drop, within other frameworks, separate actions), since the framework
itself will be invisible for the regular user. The developer should see
only the patterns that are contributed.
Scripting of operations is a think worth thinking of and we can include it
in future versions.
I hope I have answered your questions,
Dimitar
|
|
| |
| Re: Comments on Project Proposal [message #468619 is a reply to message #468617] |
Thu, 21 May 2009 22:11 |
Konstantin Komissarchik
Messages: 1077
Registered: July 2009 |
Senior Member |
|
|
As someone who has learned this the hard way (faceted project framework),
starting technology out in WTP with intentions of moving it lower in the
stack at a later date sounds good in theory, but is exceptionally costly
approach in practice.
It is significantly cheaper to start without WTP dependencies (including
dependencies on WTP namespaces (org.eclipse.jst and org.eclipse.wst)) than
it is to start out this way and then try to pull out, while maintaining
backwards compatibility for WTP. You also have to consider how you plan to
build a community around this. Again speaking from experience, it doesn't
matter how good your technology is, you will not get adoption and grow
your community beyond the sphere of adopters who build on top of WTP
already while this technology is part of WTP. Providing a separate
download will not change this situation in a meaningful way as perception
of positioning, influence and tie-ins factor in significantly in adoption
decisions. Further, you are always in danger of someone undermining all of
your work by building something similar in function lower in the stack (or
at least outside of confines of any one particular vertical tooling
project).
I would recommend that you guys at least discuss this with the Platform
team. I appreciate that you already have a code base built on the data
model framework, but if that's the only reason that you can't start out
outside of WTP then you are selling your technology short. In the long
run, it would be better for you to invest the time to remove your
dependency on the data model framework so that you can start immediately
positioned to maximize your potential community.
- Konstantin
|
|
|
| Re: Comments on Project Proposal [message #468621 is a reply to message #468619] |
Sat, 23 May 2009 02:43 |
Boris Bokowski
Messages: 272
Registered: July 2009 |
Senior Member |
|
|
Hi Dimitar,
Not sure where in the Platform this may end up but how about a call before
the creation review to get to know each other, and what you are trying to
accomplish? If the proposed technology is already implemented in a product,
you could even give a demo of it. Just set up a poll on doodle.ch with
suggested times...
Boris
"Konstantin Komissarchik" <konstantin.komissarchik@oracle.com> wrote in
message news:4e80f83e0e86aee8b6da48456251f61e$1@www.eclipse.org...
> As someone who has learned this the hard way (faceted project framework),
> starting technology out in WTP with intentions of moving it lower in the
> stack at a later date sounds good in theory, but is exceptionally costly
> approach in practice.
> It is significantly cheaper to start without WTP dependencies (including
> dependencies on WTP namespaces (org.eclipse.jst and org.eclipse.wst)) than
> it is to start out this way and then try to pull out, while maintaining
> backwards compatibility for WTP. You also have to consider how you plan to
> build a community around this. Again speaking from experience, it doesn't
> matter how good your technology is, you will not get adoption and grow
> your community beyond the sphere of adopters who build on top of WTP
> already while this technology is part of WTP. Providing a separate
> download will not change this situation in a meaningful way as perception
> of positioning, influence and tie-ins factor in significantly in adoption
> decisions. Further, you are always in danger of someone undermining all of
> your work by building something similar in function lower in the stack (or
> at least outside of confines of any one particular vertical tooling
> project).
> I would recommend that you guys at least discuss this with the Platform
> team. I appreciate that you already have a code base built on the data
> model framework, but if that's the only reason that you can't start out
> outside of WTP then you are selling your technology short. In the long
> run, it would be better for you to invest the time to remove your
> dependency on the data model framework so that you can start immediately
> positioned to maximize your potential community.
>
> - Konstantin
>
|
|
| | | | | | |
| Re: Comments on Project Proposal [message #560602 is a reply to message #468613] |
Wed, 20 May 2009 07:18 |
Dimitar Giormov
Messages: 13
Registered: July 2009 |
Junior Member |
|
|
Hi Konstantin,
First we absolutely agree that the framework should in the end go in the
platform. But we also think that the WTP Data Model framework should go
there. The PAVE framework depends only on the data model framework from
WTP, so we intend to create a separate downloadable with DMF as a first
step.
And if the PAVE framework gets enough attention we can try to push it in
the platform.
I will try to explain our vision for the framework, thus I think answering
most of your questions.
The idea behind the framework is actually easy sequencing of operations
existing and non-existing ones. We want to build a data model with all of
the data needed for all operations and then execute them in a single pass.
Also UI should be reusable, one and the same page could be used for
gathering data for creating a java class for instance and in the same time
to be reused in one of the steps of a certain pattern. Reusability is one
of our top priorities.
We want these operations to be able to be executed in headless mode, so
there can be multiple actions for executing one and the same pattern (drag
and drop, within other frameworks, separate actions), since the framework
itself will be invisible for the regular user. The developer should see
only the patterns that are contributed.
Scripting of operations is a think worth thinking of and we can include it
in future versions.
I hope I have answered your questions,
Dimitar
|
|
| |
| Re: Comments on Project Proposal [message #560616 is a reply to message #468617] |
Thu, 21 May 2009 22:11 |
Konstantin Komissarchik
Messages: 1077
Registered: July 2009 |
Senior Member |
|
|
As someone who has learned this the hard way (faceted project framework),
starting technology out in WTP with intentions of moving it lower in the
stack at a later date sounds good in theory, but is exceptionally costly
approach in practice.
It is significantly cheaper to start without WTP dependencies (including
dependencies on WTP namespaces (org.eclipse.jst and org.eclipse.wst)) than
it is to start out this way and then try to pull out, while maintaining
backwards compatibility for WTP. You also have to consider how you plan to
build a community around this. Again speaking from experience, it doesn't
matter how good your technology is, you will not get adoption and grow
your community beyond the sphere of adopters who build on top of WTP
already while this technology is part of WTP. Providing a separate
download will not change this situation in a meaningful way as perception
of positioning, influence and tie-ins factor in significantly in adoption
decisions. Further, you are always in danger of someone undermining all of
your work by building something similar in function lower in the stack (or
at least outside of confines of any one particular vertical tooling
project).
I would recommend that you guys at least discuss this with the Platform
team. I appreciate that you already have a code base built on the data
model framework, but if that's the only reason that you can't start out
outside of WTP then you are selling your technology short. In the long
run, it would be better for you to invest the time to remove your
dependency on the data model framework so that you can start immediately
positioned to maximize your potential community.
- Konstantin
|
|
|
| Re: Comments on Project Proposal [message #560623 is a reply to message #468619] |
Sat, 23 May 2009 02:43 |
Boris Bokowski
Messages: 272
Registered: July 2009 |
Senior Member |
|
|
Hi Dimitar,
Not sure where in the Platform this may end up but how about a call before
the creation review to get to know each other, and what you are trying to
accomplish? If the proposed technology is already implemented in a product,
you could even give a demo of it. Just set up a poll on doodle.ch with
suggested times...
Boris
"Konstantin Komissarchik" <konstantin.komissarchik@oracle.com> wrote in
message news:4e80f83e0e86aee8b6da48456251f61e$1@www.eclipse.org...
> As someone who has learned this the hard way (faceted project framework),
> starting technology out in WTP with intentions of moving it lower in the
> stack at a later date sounds good in theory, but is exceptionally costly
> approach in practice.
> It is significantly cheaper to start without WTP dependencies (including
> dependencies on WTP namespaces (org.eclipse.jst and org.eclipse.wst)) than
> it is to start out this way and then try to pull out, while maintaining
> backwards compatibility for WTP. You also have to consider how you plan to
> build a community around this. Again speaking from experience, it doesn't
> matter how good your technology is, you will not get adoption and grow
> your community beyond the sphere of adopters who build on top of WTP
> already while this technology is part of WTP. Providing a separate
> download will not change this situation in a meaningful way as perception
> of positioning, influence and tie-ins factor in significantly in adoption
> decisions. Further, you are always in danger of someone undermining all of
> your work by building something similar in function lower in the stack (or
> at least outside of confines of any one particular vertical tooling
> project).
> I would recommend that you guys at least discuss this with the Platform
> team. I appreciate that you already have a code base built on the data
> model framework, but if that's the only reason that you can't start out
> outside of WTP then you are selling your technology short. In the long
> run, it would be better for you to invest the time to remove your
> dependency on the data model framework so that you can start immediately
> positioned to maximize your potential community.
>
> - Konstantin
>
|
|
| | | | | | |
Goto Forum:
Current Time: Thu Apr 25 06:29:49 GMT 2024
Powered by FUDForum. Page generated in 0.04429 seconds |
] 
Search
Help
Register
Login
Home
