Re: [suggestion] - OSGI WebConsole

2011-03-03 Thread Felix Meschberger
Hi,

First: This is the last cross posted message from me. Future
communication should and will take place on the Felix list.

Please, come to that list and discuss with us here !

Am Donnerstag, den 03.03.2011, 07:40 + schrieb Charles Moulliard: 
 On Wed, Mar 2, 2011 at 11:43 PM, Felix Meschberger fmesc...@adobe.com wrote:
  Hi,
 
  As the original developer of the Web Console I am flattered by this
  activity ;-)
 
  Am Mittwoch, den 02.03.2011, 17:37 + schrieb Charles Moulliard:
  Hi,
 
  As three users of Apache Felix WebConsole project, I contact you to
  have your opinion regarding to frameworks (JSon, javascript, ...)
  usage made with Apache Felix WebConsole. The project has been made
  light to avoid dependencies with external librairies but the way that
  it is used today lack of structure, complicate the
 
  Why lack of structure ?
 
  Because by example we fill data (for a table) in different java methods 
  and render it using also different javascripts function. This is really 
  paintful to rebuild the global picture regarding what we see in the html 
  (code source) of the browser with where in the code, the different parts 
  are prepare and assemble.

Well, this is just an implementation detail of the plugins coming with
the Web Console proper. Nowhere is it prescribed that you have to do it
this way.

I think Guillaume has clearly described the issues around this.

 
 
  development of screens and decrease development productivity, html
  code is mixed in javascript, json variables are set everywhere in the
  code and use in several of javascript functions, no template is used
  to render html pages, locale is not used by all of us to translate
  text, 
 
  That's all not required. The simplest plugin is a Servlet service with
  two service properties. Nothing fancy, really ;-)
 
 
  The consequence of that is that some developers are very frustrated
  and would like to make some suggestions about Webconsole. Within Karaf
  community we already started this discussion and now we would like to
  share with you some ideas about the future roadmap of Felix
  WebConsole.
 
  I'm sorry you are very frustrated .. at the same time I do not
  understand why you don't come to the Felix dev or user list to ask
  questions ? This is a bit frustrating to me.
 
 
  The word frustrated was certainly too excessive and in fact, the 
  discussion that we have started here is a good starting point to 
  collect/gather opinions and improve the existing situation.

The Web Console lives in the Apache Felix project. So please, come to
use and talk to us. Start the discussion here and with us. We will
listen to you.

 
 
  Here are the different scenario possible :
 
  (1) Improve the existing usage of the frameworks JSon, OSGI and
  javascript by defining/providing a template project containing dummy
  code + guidelines/best practices to develop properly and so
  improve/increase productivity. This could be done with an archetype
 
  Sounds like an excellent idea.
 
 
  (2) Switch the existing architectural model to use frameworks like
  Apache Wicket or Vaadin where the content is clearly separated from
  the server side code. Apache Wicket and Vaadin librairies are already
  osgified so their integration in project like karaf, sling or felix
  will be done seamless even if we have the overhead to deploy them. But
  this is also the same for bundles like PAX-Web, ...
 
  Sounds like not soo go an idea. There is a reason why we don't use any
  of these frameworks: We wanted (and still want) to minimize
  dependencies. Adding such a nice GUI framework to the game makes it more
  complicated (remember the simple Servlet service mentioned above ?) and
  heavy weight.
 
  Before to take a definitive decision, I propose that we investigate the 
  possibilities offered by Apache Wicket, PAX-Wicket or even Vaadin using a 
  few rendering components and analyse if the approach suggested is worse 
  or can provide enhancements for the existing platform without too much 
  impact.

I know very well about the functionality provided by those platforms.
But all these platforms come with additional dependencies. And this is
where I start placing question marks.

Also be reminded that Web Console is currently being used in embedded
platforms which don't have place to spare for the developers delight.

 
 
  However, this is not to say, that we should not explore ways on how
  mechanism like JSP (or generic server side scripting) etc. can be used.
 
 
  No matter which scenario we will decide to adopt, we could also create
  a project to develop all together the OSGI WebConsole used by our
  projects and promote it as a new Apache project -- Name suggested
  Apache Orion. The scope of this project could be extended to include
  additional management, registration of datasources, 
 
  What would be the advantage of developing the Web Console in its own
  TLP ? Is there something missing when living in the 

[suggestion] - OSGI WebConsole

2011-03-02 Thread Charles Moulliard
Hi,

As three users of Apache Felix WebConsole project, I contact you to
have your opinion regarding to frameworks (JSon, javascript, ...)
usage made with Apache Felix WebConsole. The project has been made
light to avoid dependencies with external librairies but the way that
it is used today lack of structure, complicate the
development of screens and decrease development productivity, html
code is mixed in javascript, json variables are set everywhere in the
code and use in several of javascript functions, no template is used
to render html pages, locale is not used by all of us to translate
text, 

The consequence of that is that some developers are very frustrated
and would like to make some suggestions about Webconsole. Within Karaf
community we already started this discussion and now we would like to
share with you some ideas about the future roadmap of Felix
WebConsole.

Here are the different scenario possible :

(1) Improve the existing usage of the frameworks JSon, OSGI and
javascript by defining/providing a template project containing dummy
code + guidelines/best practices to develop properly and so
improve/increase productivity. This could be done with an archetype

(2) Switch the existing architectural model to use frameworks like
Apache Wicket or Vaadin where the content is clearly separated from
the server side code. Apache Wicket and Vaadin librairies are already
osgified so their integration in project like karaf, sling or felix
will be done seamless even if we have the overhead to deploy them. But
this is also the same for bundles like PAX-Web, ...

No matter which scenario we will decide to adopt, we could also create
a project to develop all together the OSGI WebConsole used by our
projects and promote it as a new Apache project -- Name suggested
Apache Orion. The scope of this project could be extended to include
additional management, registration of datasources, 

What do you think my propositions ?

Regards,

Charles Moulliard
Apache Committer

Blog : http://cmoulliard.blogspot.com
Twitter : http://twitter.com/cmoulliard
Linkedin : http://www.linkedin.com/in/charlesmoulliard
Skype: cmoulliard


Re: [suggestion] - OSGI WebConsole

2011-03-02 Thread Felix Meschberger
Hi,

As the original developer of the Web Console I am flattered by this
activity ;-)

Am Mittwoch, den 02.03.2011, 17:37 + schrieb Charles Moulliard: 
 Hi,
 
 As three users of Apache Felix WebConsole project, I contact you to
 have your opinion regarding to frameworks (JSon, javascript, ...)
 usage made with Apache Felix WebConsole. The project has been made
 light to avoid dependencies with external librairies but the way that
 it is used today lack of structure, complicate the

Why lack of structure ?

 development of screens and decrease development productivity, html
 code is mixed in javascript, json variables are set everywhere in the
 code and use in several of javascript functions, no template is used
 to render html pages, locale is not used by all of us to translate
 text, 

That's all not required. The simplest plugin is a Servlet service with
two service properties. Nothing fancy, really ;-)

 
 The consequence of that is that some developers are very frustrated
 and would like to make some suggestions about Webconsole. Within Karaf
 community we already started this discussion and now we would like to
 share with you some ideas about the future roadmap of Felix
 WebConsole.

I'm sorry you are very frustrated .. at the same time I do not
understand why you don't come to the Felix dev or user list to ask
questions ? This is a bit frustrating to me.

 
 Here are the different scenario possible :
 
 (1) Improve the existing usage of the frameworks JSon, OSGI and
 javascript by defining/providing a template project containing dummy
 code + guidelines/best practices to develop properly and so
 improve/increase productivity. This could be done with an archetype

Sounds like an excellent idea.

 
 (2) Switch the existing architectural model to use frameworks like
 Apache Wicket or Vaadin where the content is clearly separated from
 the server side code. Apache Wicket and Vaadin librairies are already
 osgified so their integration in project like karaf, sling or felix
 will be done seamless even if we have the overhead to deploy them. But
 this is also the same for bundles like PAX-Web, ...

Sounds like not soo go an idea. There is a reason why we don't use any
of these frameworks: We wanted (and still want) to minimize
dependencies. Adding such a nice GUI framework to the game makes it more
complicated (remember the simple Servlet service mentioned above ?) and
heavy weight.

However, this is not to say, that we should not explore ways on how
mechanism like JSP (or generic server side scripting) etc. can be used.

 
 No matter which scenario we will decide to adopt, we could also create
 a project to develop all together the OSGI WebConsole used by our
 projects and promote it as a new Apache project -- Name suggested
 Apache Orion. The scope of this project could be extended to include
 additional management, registration of datasources, 

What would be the advantage of developing the Web Console in its own
TLP ? Is there something missing when living in the Felix project ?

Regards
Felix

 
 What do you think my propositions ?
 
 Regards,
 
 Charles Moulliard
 Apache Committer
 
 Blog : http://cmoulliard.blogspot.com
 Twitter : http://twitter.com/cmoulliard
 Linkedin : http://www.linkedin.com/in/charlesmoulliard
 Skype: cmoulliard




Re: [suggestion] - OSGI WebConsole

2011-03-02 Thread Charles Moulliard
On Wed, Mar 2, 2011 at 11:43 PM, Felix Meschberger fmesc...@adobe.com wrote:
 Hi,

 As the original developer of the Web Console I am flattered by this
 activity ;-)

 Am Mittwoch, den 02.03.2011, 17:37 + schrieb Charles Moulliard:
 Hi,

 As three users of Apache Felix WebConsole project, I contact you to
 have your opinion regarding to frameworks (JSon, javascript, ...)
 usage made with Apache Felix WebConsole. The project has been made
 light to avoid dependencies with external librairies but the way that
 it is used today lack of structure, complicate the

 Why lack of structure ?

 Because by example we fill data (for a table) in different java methods and 
 render it using also different javascripts function. This is really 
 paintful to rebuild the global picture regarding what we see in the html 
 (code source) of the browser with where in the code, the different parts 
 are prepare and assemble.


 development of screens and decrease development productivity, html
 code is mixed in javascript, json variables are set everywhere in the
 code and use in several of javascript functions, no template is used
 to render html pages, locale is not used by all of us to translate
 text, 

 That's all not required. The simplest plugin is a Servlet service with
 two service properties. Nothing fancy, really ;-)


 The consequence of that is that some developers are very frustrated
 and would like to make some suggestions about Webconsole. Within Karaf
 community we already started this discussion and now we would like to
 share with you some ideas about the future roadmap of Felix
 WebConsole.

 I'm sorry you are very frustrated .. at the same time I do not
 understand why you don't come to the Felix dev or user list to ask
 questions ? This is a bit frustrating to me.


 The word frustrated was certainly too excessive and in fact, the 
 discussion that we have started here is a good starting point to 
 collect/gather opinions and improve the existing situation.


 Here are the different scenario possible :

 (1) Improve the existing usage of the frameworks JSon, OSGI and
 javascript by defining/providing a template project containing dummy
 code + guidelines/best practices to develop properly and so
 improve/increase productivity. This could be done with an archetype

 Sounds like an excellent idea.


 (2) Switch the existing architectural model to use frameworks like
 Apache Wicket or Vaadin where the content is clearly separated from
 the server side code. Apache Wicket and Vaadin librairies are already
 osgified so their integration in project like karaf, sling or felix
 will be done seamless even if we have the overhead to deploy them. But
 this is also the same for bundles like PAX-Web, ...

 Sounds like not soo go an idea. There is a reason why we don't use any
 of these frameworks: We wanted (and still want) to minimize
 dependencies. Adding such a nice GUI framework to the game makes it more
 complicated (remember the simple Servlet service mentioned above ?) and
 heavy weight.

 Before to take a definitive decision, I propose that we investigate the 
 possibilities offered by Apache Wicket, PAX-Wicket or even Vaadin using a 
 few rendering components and analyse if the approach suggested is worse or 
 can provide enhancements for the existing platform without too much impact.


 However, this is not to say, that we should not explore ways on how
 mechanism like JSP (or generic server side scripting) etc. can be used.


 No matter which scenario we will decide to adopt, we could also create
 a project to develop all together the OSGI WebConsole used by our
 projects and promote it as a new Apache project -- Name suggested
 Apache Orion. The scope of this project could be extended to include
 additional management, registration of datasources, 

 What would be the advantage of developing the Web Console in its own
 TLP ? Is there something missing when living in the Felix project ?

 If we add new features/functionalities (like also mentioned by Jean 
 Baptiste D' Onofre), the debate about the location
of this project is not insignifiant and becomes relevant like also
that this project becomes a TLP. I do not at all defend a point of
view to move this Felix sub-project but open the debate


 Regards
 Felix


 What do you think my propositions ?

 Regards,

 Charles Moulliard
 Apache Committer

 Blog : http://cmoulliard.blogspot.com
 Twitter : http://twitter.com/cmoulliard
 Linkedin : http://www.linkedin.com/in/charlesmoulliard
 Skype: cmoulliard





Re: [suggestion] - OSGI WebConsole

2011-03-02 Thread Guillaume Nodet
On Thu, Mar 3, 2011 at 08:40, Charles Moulliard cmoulli...@gmail.com wrote:
 On Wed, Mar 2, 2011 at 11:43 PM, Felix Meschberger fmesc...@adobe.com wrote:
 Hi,

 As the original developer of the Web Console I am flattered by this
 activity ;-)

 Am Mittwoch, den 02.03.2011, 17:37 + schrieb Charles Moulliard:
 Hi,

 As three users of Apache Felix WebConsole project, I contact you to
 have your opinion regarding to frameworks (JSon, javascript, ...)
 usage made with Apache Felix WebConsole. The project has been made
 light to avoid dependencies with external librairies but the way that
 it is used today lack of structure, complicate the

 Why lack of structure ?

 Because by example we fill data (for a table) in different java methods 
 and render it using also different javascripts function. This is really 
 paintful to rebuild the global picture regarding what we see in the html 
 (code source) of the browser with where in the code, the different parts 
 are prepare and assemble

That's a really common thing when using AJAX to render data.  I think
you missed the point that the servlet render JSON data which could be
used by another tool / plugin with a different view.
So the data acces / json rendering is done on the server side while
the html pages usually use javascript to render the json data into
html.  That's a clean separation and allows reloading the page without
reloading the full html page.

.


 development of screens and decrease development productivity, html
 code is mixed in javascript, json variables are set everywhere in the
 code and use in several of javascript functions, no template is used
 to render html pages, locale is not used by all of us to translate
 text, 

 That's all not required. The simplest plugin is a Servlet service with
 two service properties. Nothing fancy, really ;-)


 The consequence of that is that some developers are very frustrated
 and would like to make some suggestions about Webconsole. Within Karaf
 community we already started this discussion and now we would like to
 share with you some ideas about the future roadmap of Felix
 WebConsole.

 I'm sorry you are very frustrated .. at the same time I do not
 understand why you don't come to the Felix dev or user list to ask
 questions ? This is a bit frustrating to me.


 The word frustrated was certainly too excessive and in fact, the 
 discussion that we have started here is a good starting point to 
 collect/gather opinions and improve the existing situation.




 Here are the different scenario possible :

 (1) Improve the existing usage of the frameworks JSon, OSGI and
 javascript by defining/providing a template project containing dummy
 code + guidelines/best practices to develop properly and so
 improve/increase productivity. This could be done with an archetype

 Sounds like an excellent idea.


 (2) Switch the existing architectural model to use frameworks like
 Apache Wicket or Vaadin where the content is clearly separated from
 the server side code. Apache Wicket and Vaadin librairies are already
 osgified so their integration in project like karaf, sling or felix
 will be done seamless even if we have the overhead to deploy them. But
 this is also the same for bundles like PAX-Web, ...

 Sounds like not soo go an idea. There is a reason why we don't use any
 of these frameworks: We wanted (and still want) to minimize
 dependencies. Adding such a nice GUI framework to the game makes it more
 complicated (remember the simple Servlet service mentioned above ?) and
 heavy weight.

 Before to take a definitive decision, I propose that we investigate the 
 possibilities offered by Apache Wicket, PAX-Wicket or even Vaadin using a 
 few rendering components and analyse if the approach suggested is worse or 
 can provide enhancements for the existing platform without too much impact.


 However, this is not to say, that we should not explore ways on how
 mechanism like JSP (or generic server side scripting) etc. can be used.


 No matter which scenario we will decide to adopt, we could also create
 a project to develop all together the OSGI WebConsole used by our
 projects and promote it as a new Apache project -- Name suggested
 Apache Orion. The scope of this project could be extended to include
 additional management, registration of datasources, 

 What would be the advantage of developing the Web Console in its own
 TLP ? Is there something missing when living in the Felix project ?

 If we add new features/functionalities (like also mentioned by Jean 
 Baptiste D' Onofre), the debate about the location
 of this project is not insignifiant and becomes relevant like also
 that this project becomes a TLP. I do not at all defend a point of
 view to move this Felix sub-project but open the debate


 Regards
 Felix


 What do you think my propositions ?

 Regards,

 Charles Moulliard
 Apache Committer

 Blog : http://cmoulliard.blogspot.com
 Twitter : http://twitter.com/cmoulliard
 Linkedin :