So far as I am concerned, the basic ideas in Struts are superior.
Struts is poor architected in many ways and needs to be improved.
Introducing basic interfaces and decoupling so far as is practicable are
just two necessities. However, the ideal, in my opinion, would be to
improve Struts and to provide "connected" technologies like a
sophisticated view, etc. JSF can go where it wants in my view and I
wish people who like it the best. I just cannot endorse the basic
concepts myself and believe they can never become popular. I really
don't think JSF is for the better.
Sean Schofield wrote:
Shale very deliberately presumes the presence of JSF as a foundation
technology. That means that, among other things, Shale does not need
to reinvent a bunch of technology that JSF already provides (in
particular, managed beans, page navigation, the request processing
lifecycle for form submits, and value/method binding expressions).
Ironically, Shale itself doesn't care a lot about which actual JSF
components you are using :-). It wants JSF for its framework
capabilities. In turn, this lets the development of Shale focus on
the areas that JSF does not, or does not yet, cover.
Good point. Perhaps some of the confusion is over the lack of
understanding of what JSF does. I did not really understand the Shale
proposal until I did a lot of reading on JSF.
So far, however, the Struts developers have been unwilling to make the
leap to "assume JSF as a base technology, then build on top" -- the
very assumption that is the foundation to the whole idea.
IMO this calls for abandoning the notion of Struts as we know it.
Developers like myself have come to love Struts for how it helped
simplify and standardize web applications. We've grown attached to
it, so there is reluctance to abandon this approach in favor of a new
one.
You've mentioned in previous posts about how JSF has the advantage of
being designed from scratch with all of the knowledge gained from the
work on Struts (and other frameworks.) I agree that at some point we
will have to say good bye to our beloved Struts. Also, there will be
a lengthy transition where development on Struts continues and even
longer transition where people will continue to use it. Many "late
adopters" are just now getting around to learning Struts. We're
talking years until they get to JSF.
I would propose we consider halting major development efforts on
Struts after the 1.3 release. The 1.3 release will use commons-chain
for request processing which will give even more flexability to Struts
users. After that release though, it would seem that Struts would be
all grown up and ready to leave the nest. Yes we could always add
something, but why not be bold and try something new (as Craig is
suggesting)?
One suggestion would be to not really consider Shale as Struts 2.0.
From what I understand, it really has very little to do with Struts as
we know it. Ultimately we should consider releasing it as its own
project with a new name (Shale or whatever.)
Just some thoughts.
Craig McClanahan
sean
