Zoran Avtarovski wrote:
I thought it looks pretty good, I think projects like this have their place.
Especially, in light of the fact that we're not going to see Action2 for a
little while (even then there'll be some tweaking to get it right)
Well, the fact is that Action2 is just Webwork. What you see on the dev
list is just a lot of discussion of renaming packages and classes and
taglib prefixes, but hardly anything substantive.
And Webwork is hardly something you need to wait for, it can be used
now. Well, not only can it be used now, it has been in a very solidly
usable state for years. After all, this is the framework underlying
products like Jive forums, Confluence, JIRA, and so on. AFAICS, the only
thing you would have to do to move from Webwork 2.2 to Struts Action 2
is do a bunch of package/class renaming on your code.
AFAICS, if you think that you will eventually be using Struts Action 2,
you might as well start using Webwork now -- at least for new codebases.
and there
are a lot of Struts apps out there that want to take advantage of Java 5.
I get the impression people are giving up on Struts (Struts action 1, that
is).
Well, if the Struts developers themselves are saying that Struts 1.x is
technically obsolete and is not going to be the focus of further
development effort, AFAICS, it is silly not to just take this at face value.
As an outside observer, the problem I see is that they are not providing
much clarity about what existing users should do. They have this idea
that they'll offer two frameworks under the "Struts umbrella" that
embody two completely different approaches, Struts Action 2 (Webwork)
and Shale (based on JSF). And then at the overall Struts project level,
I guess they intend to maintain some kind of official agnosticism as to
which approach is to be preferred -- rather than simply offering their
own recommended transition guidelines.
But again, if the Struts developers are themselves saying that Struts
1.x is technically obsolete, I think this should be taken at face value.
This has clear implications and it is silly to fight against this. Also,
you know, this business of people flaming me, for example, basically for
just recapitulating what the Struts developers are themselves saying --
this is really the height of silliness.
I see that advantages in other frameworks but most of these frameworks
were around when people made the decision to base their projects on struts.
Well, Struts had huge placement/visibility advantages over competitors
such as Webwork. Your above point would be more cogent if these things
really had been competing on a level playing field. (But they weren't of
course...)
Which leads me to believe that a lot of people, like myself, like using
struts. I'm not naïve, I know it has issues, but it also has advantages and
at the end of the day we have too much invested to just throw it all out and
start from scratch using something else.
Well, I don't know. Out of curiosity, have you carried out an appraisal
of how much work it is to transition an app from Struts Action 1 to
Action 2 (i.e. Webwork)?
I ask because I honestly don't know. Though one thing is clear enough
IMO: if you are going to eventually assume that migration cost, whatever
it is, you might as well assume it sooner rather than later, since the
cost is the same and you might as well get the benefits of that
migration sooner rather than later.
I don't want to start the sort of crap we had a few weeks ago, so I'm going
to try and limit the scope of my question.
Well, if you pose a perfectly legitimate question of greater scope and
this leads to extreme nastiness and people personally attacking you,
what does this mean? Basically, I dunno... if people won't maintain
basic rules of debate and discourse, and when you make legitimate
points, and raise legitimate questions, are going to start screaming,
well what is one to do?
Again, I ask because I honestly don't know. If certain completely
legitimate questions are "taboo", then how is there much basis for
getting anywhere?
I like that there are now three
frameworks under the struts umbrella and I think each serves a purpose, but
what I'm concerned about is that struts 1 will be left in legacy mode where
you might get the odd bug fix but no forward progress.
Yes, but Zoran, you have been told that this is the case. So what do you
plan to do? (I hope your "plan" isn't the other people's plan, which is
to flame *me* when I repeat what the Struts developers themselves have
been saying....)
As I said we have way
too much invested in struts 1 projects and nobody here wants to jump in and
start rebuilding each project using a new framework. We would rather have
the ability to take advantage of the new features and developments in
extending and maintaining these projects without having to rebuild the whole
project.
In essence what I like about Strecks is that it enables us to:
1. Take advantage of Java 5 features
2. Take advantage of simplified development structure
3. Deal with the worst bits of Struts while still maintaining backwards
compatibility.
Is this something we are going to see as part of ongoing struts action
development?
My sense of things is that Strecks is another person's third-party
effort. It is not part of the Struts project. Any new development you
can expect from the Struts developers is bifurcating between Action2
(Webwork) and Shale.
I would venture to guess, just as an outside observer, that if the
author of Strecks is not given commit access to Struts itself, then he
may run into limitations in the Struts codebase and end up making
modifications to support things that Strecks needs, and end up with a
forked version of the Struts 1.x code.
Of course, logically, if this guy is interested in developing further on
top of Struts 1.x, and existing Struts committers aren't, he probably
should be let in and Strecks might as well become Struts 1.3.x. or
Struts 1.4.x and so on.
Somehow, I don't think that's likely to happen though. *Probably* (note
that I am speculating because I don't speak for others) the preference
of the Struts people would be just to let the Struts 1.x codebase more
or less rot and encourage people to move to Action2 or Shale.
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
Z.
http://www.theserverside.com/news/thread.tss?thread_id=39840
.V
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
