Re: RFC: BLOBAction and Struts Bloat

2004-09-21 Thread Craig McClanahan
Thanks for your comments, Michael.  If you've been following the dev
list lately, you've seen some beginning discussions on a Struts 2.0
rearchitecting that would indeed leverage everything we've all learned
in the four years since Struts was first created.  I have some
specific proposals to make in this regard, which I'll be sharing when
I return from an extended trip next week.

That being said, one of the factors that has made Struts so popular is
a commitment to take care of existing users.  It would be somewhat
irresponsible for us to completely stopping development of the Struts
1.x architecture, or just doing bug fixes.  Therefore, we need to do a
1.3 release in the interim time period, focused on a small number of
changes, such as:

* Changing base API platform to Servlet 2.3 and JSP 1.2,
  so Struts apps can count on things like filters and event
  listeners.

* Refactoring the RequestProcessor class to use Commons Chain
  (based on the code in the contrib/struts-chain) that supports the
  1.2 request processing lifecycle semantics, but is more easily
  customized than the current architecture.  You'll also be able to
  use the Chain paradigm for your own business logic if you like.

* Provide a second request processor implementation chain that
  operates in a portlet (JSR-168) environment.

* Split the Struts monolithic release into separate releases of the
  core framework, the tag libraries, the examples, and so on.
  This will help us accelerate the turnaround of releases.

In the near future, you'll also see the initial release candidate of
the Struts-Faces integration library (packaged separately from the
rest of Struts) that allows JavaServer Faces to be used with Struts
1.1 or 1.2 based applications, including the use of the Tiles
Framework and the Validator Framework.

Note that I do *not* see any of the developers interested in
continuing the development of the Struts HTML tag libraries, as other
view tier choices (like JSF) are becoming available.

Craig

PS:  With regards to migrating to SVN (commented on in one of the
replies), doing both 1.3 and 2.0 together on SVN will be vastly more
productive than using CVS for 1.3 and SVN for 2.0.

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



RE: RFC: BLOBAction and Struts Bloat

2004-09-21 Thread Matthias Wessendorf
Craig,

 In the near future, you'll also see the initial release 
 candidate of the Struts-Faces integration library (packaged 
 separately from the rest of Struts) that allows JavaServer 
 Faces to be used with Struts 1.1 or 1.2 based applications, 
 including the use of the Tiles Framework and the Validator Framework.

will the Struts-Faces integration library be a subproject of Apache
Struts?
Like http://struts.apache.org/struts-faces for instance?
Or what are your plans on it?


 Note that I do *not* see any of the developers interested in 
 continuing the development of the Struts HTML tag libraries, 
 as other view tier choices (like JSF) are becoming available.


Btw. like Michael, I am interessted in your proposals on
Struts 2.0 too :)

Matthias

 Craig
 
 PS:  With regards to migrating to SVN (commented on in one of 
 the replies), doing both 1.3 and 2.0 together on SVN will be 
 vastly more productive than using CVS for 1.3 and SVN for 2.0.
 
 -
 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]



Re: RFC: BLOBAction and Struts Bloat

2004-09-21 Thread Vic
Matthias Wessendorf wrote:
Craig,

Btw. like Michael, I am interessted in your proposals on
Struts 2.0 too :)


Checkout this:
http://cvs.apache.org/viewcvs.cgi/jakarta-struts/contrib/struts-chain
.V
--
Please post on Rich Internet Applications User Interface (RiA/SoA)
http://www.portalvu.com
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


Re: RFC: BLOBAction and Struts Bloat

2004-09-20 Thread bryan
There are some things that I certainly agree with michael on and this 
is one of them.

I think ( and i'm not alone in thinking this ) that struts needs a complete 
overhaul. 

At present the only reason I use it is because all the existing tooling 
supports it. 

eg m7 nitrox, WSAD, MyEclipse etc etc etc 

If the spring tooling was as mature as this I would use it instead. 

The project has a reputation for inflexibility, complexity and a lot of 
unnecessary inheritance. 

These are all hangovers from EJB with it's multiple interfaces and 
fixed inheritance approach with the associated difficulties of testing,
slow development cycle etc. 

Struts has done a superb job in the past otherwise it wouldn't be
so popular and there is a fantastic pool of talent in the struts 
development team but it does seem like you are loosing mind-share.

If this were not the case there wouldn't be so many copy cat 
web MVC frameworks springing up.

Perhaps it is time for an overhaul of struts. Rather than moving 
your codebase to subversion perhaps you should leave your 1.* 
development on CVS and start a completely new attempt for your 
2.* series on a fresh server. 

I do think that struts is a great product and I am a big fan of it's tag
libs and the MVC approach in general but I do think that it may 
be time for a rethink 

Everything is going POJO, attributes are the way forward. 1.5 will
mean that collections are no longer unknown bundles of objects.

A lot of people also want AOP features like declaritive behaviour 
and easier testing. 

People also want things to be more flexible and have shorter development
cycles. 

I've been a developer for 6 years now. Programmed in about 5 languages
and still I find the slowest part of the development cycle is handling the 
presentation layer with it's multiple reloads etc. 

I mean i could write a desktop application now with remoting to do the 
same job quicker than i could write the web interface. 

That doesn't seem to be the promise offered when people first started 
developing web applications. And I remember that far back.

One thing that I am certainly not suggesting and I want to make this 
clear is that I am not suggesting that anyone fork the codebase.

There is certainly too much clutter surrounding  the project at present.

It's better if you make it easier to do the difficult stuff rather than vice 
versa and having everything but the kitchen sink thrown in there doesn't
make it any easier.

--b







approach 


On Mon, 20 Sep 2004 06:31:24 -0700, Michael McGrady
[EMAIL PROTECTED] wrote:
 Open source and painting are connected in someways.  The biggest mistake
 of the amateur painter is to keep adding colors when the painting is
 done.  The result is always that murky, dark, ugly, look.  Likewise,
 open source, filled with people with egos, sometimes does not know when
 something is done, finished, damed good and certainly good enough.
 
 I have a real concern that Struts is going to continue to be bloated
 with what are not Struts, not part of the framework, but what are
 instead really very useful uses of Struts.  DispatchAction and its
 progeny are just one example of this.  I would suggest that such
 offerings would better be managed in a separate application called
 something like Struts Patterns or Struts Best Practices.  Perhaps
 such classes could be released as classes and either maintained (or
 preferrably not) independent of Struts?
 
 I would hope that by the term bloat I don't convey that such classes
 are not wonderful ideas and their authors are stellar citizens.  It just
 seems to me that a framework ought to be maintained as a FRAMEWORK and
 that allowing USES OF THE FRAMEWORK to become part of the framework is a
 groundwater mistake.  Generally, I think it might be better if the
 actions package were distributed outside Struts proper.
 
 As a framework, there comes a time when something like Struts is done,
 and we need to move on to other things except diligent maintainence and
 creating clever uses.  My present predeliction is to save what presently
 exists, clean it up by jettisoned what is not needed, and to watch for
 good ideas that might arise in uses.  This is just a thought.
 
 There needs to be, I think, a clear separation of Struts patterns or
 clever uses of Struts and Struts itself.  I would suggest that something
 formal be instituted.  Thanks for your ears/eyes on this.
 
 Michael McGrady
 
 -
 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]