Nobody is forced to use HiveMind. It's there. Tapestry uses it to coordinate
the many moving parts of its infrastructure. It's convienient for
applications, or they can use whatever they want. The new, higher level of
expressiveness has allowed Tapestry to move forward in deep and subtle ways
It's allowed me to get code tested that wasn't testable in the past. Spring
does not have the level of expressiveness needed by Tapestry.

I don't understand why people will subscribe to an IoC container for thier
applications (which are often relatively trivial) and not for an underlying
framework, especially one with the internal complexity and extensibility of
Tapestry. 3.0, lacking an IoC container, was largely tapped out in terms of
extending it, and fixing the issues many people really struggle with
(parameter direction and so forth).

My Mantra is "Less Is More". I want less code, less XML, less HTML. 4.0 is
headed in the right direction by that measure. Simplcitiy, Consistency,
Efficiency, Feedback. 4.0 is an improvement on 3.0 in terms of all four
principles. Refactoring around HiveMind is largely the reason.

If you want whizz bang components right now, go write some. Tapestry has had
the necessary support, server-side and client-side, including dynamic
JavaScript, for quite some time now (2 - 3 years). There are limits to what
can be packaged with Tapestry proper, due to licensing, so I think the
really interesting (Ajax) components are always going to be a seperate
project.

The "take my ball and go home" argument doesn't carry much weight either.
I'd really prefer to see some encouragment here. I've invested years of my
life, and the equivalent of $200,000+ in lost wages, into Tapestry. When I
see messages that accuse me of some kind of intellectual self-indulgence, it
can be pretty de-motivating.

5/6/05, Gregg D Bolinger <gthought (AT) gmail (DOT) comwrote:
I'll add my 2 cents here. A lot of good points either way you look at
it in this thread.

I'd argue that what draws people to ASP.NET <http://ASP.NETis not the
components. It
is the speed in which you can click out a web application in VS.NET<http://VS.NET>

[sidetrack] I hear a lot of people rant and rave about VB.NET<http://VB.NET>because
of how fast you can build GUI's. Well, it's not VB the language that
makes that possible. It's VS.NET <http://VS.NET>. [/sidetrack].

So I think the component argument should be aimed at JSF, Wicket, etc
because in all honestly I don't believe Tapestry is competing against
.NET.

I personally like Tapestry. And I am all for change as long as it is
for the better. Don't take one step foward and 2 steps back just to
make the framework *cooler*.

My biggest complaint about Tapestry 4.0, or whatever it will be
called, is that it is integrated with HiveMind. These are 2 different
project solving 2 different problems. I don't think an IoC container
should be in the same distribution as a Web App Framework. With that
being said, I don't know how Howard is planning on using them
together. I haven't had a chance to look at any of the new Tapestry
stuff yet.

I am hoping that IoC still remains a developers option when using
Tapestry. How hard will it be to use Spring instead of HiveMind if I
want? Is that even possible?

BTW - I've taken a look at Wicket and while it does look like a
Tapestry clone, it is coming along quite nicely. If I am forced to
use HiveMind with the new Tapestry, I'll use Wicket. :)

Gregg Bolinger

5/6/05, Karthik Abram <karthik.abram (AT) neovera (DOT) comwrote:

I've not seen one blog/article/email that convinces me HiveMind provides
any
capabilities to the component writer. I believe there are changes being
made
to simplify writing components a bit, but still, they don't require
HiveMind. HiveMind will also hamper the adoption of Tapestry to some
extent.
Corporations now have to accept another SS into their support
structure. If
a HiveMind like framework is required, then why no Spring which is
clearly
more popular?

I agree with Patrick - there are lots of features (e.g. rewind cycle)
that
can be worked on. The documentation - which is outdated (including the
book)
can be brought up-to-speed. The famed Tapestry learning curve can be
addressed with good examples

In the end, SS is also a player in the free market. I think it is
telling
that Tapestry is technically a better framework, but has so small a
marketshare. Not the first time I've seen this.

Message
From: Brian K. Wallace [mailto:brian (AT) transmorphix (DOT) com]
Sent: Friday, May 06, 2005 3:01 AM
To: Tapestry users
Subject: Re: Components to fry JSF and all other frameworks

PGP SIGNED MESSAGE
Hash: SHA1

While I understand the sentiment you present below, I must disagree.
You, as a user, *should* have the loudest vote. Not that all of every
user's wants and desires will ever be filled, but the users are the
community. If you want something, speak up. If you get a lot of other
users saying it's a bad idea - well, whether it is or not it probably
won't get voted to the top of the heap of changes. Still doesn't mean it
can't end up in the codebase, just means you'd probably have to be the
one to step outside the 'user' role and implement it. In emphasizing
'should' above I acknowledge that there is usually a guiding direction
(usually being more often than not in an ideal world) that may either
eliminate or modify the needs or wants presented in a current version.
However I do not believe that the Tapestry community is just the core
that brought it into Apache and under Jakarta. There is still a sense of
directing everything toward Howard for 'approval', but I see this as
more of a case of him providing an overall direction and no real
objections presented to that direction. In the early days I did not
quite feel that (it was Howard's baby and he knew it better than
anyone), but as the community has grown and more people are utilizing
*JIRA* (note: Jira is not just for bugs) and taking Tapestry outside the
box in which it had previously been implemented it appears as though
more and more the users are powering the development. If not powering,
at least empowering. (please disregard any words that appear marketing
in nature ;-))

All that said, yes - the committers do have the only votes that _really_
count - as in "binding". However, I have yet to see a -1 from anyone
participating in a vote - binding or not - be ignored. You want the
kitchen sink before you'll change your -1? Alright - you'd probably be
vetoed. But the key to this, and any other project, is the path forward
And all the committers, while having their own views (which may or may
not match anyone else's - not even Howard's), have shown their
willingness to discuss changes.

As to the Hivemind dependency being an 'added' layer - if you look at
the core of what Tapestry requires, both pre and post Hivemind change,
you'll see that there are no truly new requirements of users. You can
still do what you've been doing and achieve the same results.
Refactoring the code to separate more of what Hivemind does best into
the Hivemind codebase and pointing Tapestry to use it doesn't add a
whole lot more to what Tapestry requires and does add a whole lot more
to what Tapestry can provide. If I write no log statements, I still
depend on logging. If I use no Hivemind integration in Tapestry, I still
depend on it. Same concept. If it makes it so developers (not speaking
of commiters here) can more easily create the cool components - I'm all
for it. And if it makes it easier for developers (speaking Tapestry
developers here) to stay flexible in the future direction of Tapestry -
I'm all for that as well.

My .02

Patrick Casey wrote:
| course I'm not a tapestry dev; I'm a user, so what I *want* to
| happen may not match what the community as a whole wants to happen.
| Likewise, the devs are the only vote that really counts :).
|
| Pat
PGP SIGNATURE
Version: GnuPG v1.2.5 (MingW32)

4PToslltH7nku5yMm7rek4k=
=0Hbg
PGP SIGNATURE

To unsubscribe, e-mail: tapestry-user-unsubscribe (AT) jakarta (DOT) apache.org
For additional commands, e-mail: tapestry-user-help (AT) jakarta (DOT) apache.org

To unsubscribe, e-mail: tapestry-user-unsubscribe (AT) jakarta (DOT) apache.org
For additional commands, e-mail: tapestry-user-help (AT) jakarta (DOT) apache.org
--

To unsubscribe, e-mail: tapestry-user-unsubscribe (AT) jakarta (DOT) apache.org
For additional commands, e-mail: tapestry-user-help (AT) jakarta (DOT) apache.org

Message
From: "Gregg D Bolinger" <gthought (AT) gmail (DOT) com>
To: "Tapestry users" <tapestry-user (AT) jakarta (DOT) apache.org>
Sent: Friday, May 06, 2005 5:23 PM
Subject: Re: Components to fry JSF and all other frameworks

I am hoping that IoC still remains a developers option when using
Tapestry. How hard will it be to use Spring instead of HiveMind if I
want? Is that even possible?

Strangely, how come I never hear anyone complaining to Rod Johnson for how
they want to use Spring MVC as web layer *but* without having Spring IoC
container around, because they only want HiveMind for that, not both ?
Rod would probably be puzzled by this strange need no less then Howard

Let me get this straight because I heard this unreasonable complaint before
about having 2 IoC containers - There is almost always some kind of
container inside some project!
Tapestry 3.0 has it also - in a form of .application file, Extensions,
EngineServices etc It is the way hw Tapestry 3.0 is built. It was blended
with whole framework so you never considered it as something separated, with
some special name. The only difference now with Tapestry 4.0 is that this
container is separated as new project (HiveMind), and it's something that
Tapestry is built on, same as before, just with difference of being separate
project, with a name.

You should not care how Tapestry is built, you should pick whatever you want
in your project. If it's Spring, so be it. Use it as you always did. The
only thing desirable for IoC frameworks is that they know how to inject
dependencies from one to another, so that wiring of dependencies would be
outside of code, and Tapestry 4.0 offers that in a form of "inject" tag.
-Vjeran

Howard Lewis Ship wrote:

>The "take my ball and go home" argument doesn't carry much weight either.
>I'd really prefer to see some encouragment here. I've invested years of my
>life, and the equivalent of $200,000+ in lost wages, into Tapestry. When I
>see messages that accuse me of some kind of intellectual self-indulgence, it
>can be pretty de-motivating.

Howard,

your work - as well as the work everyone else put into tapestry and
hivemind - is truly inspirational, at least by my standards.
I would think that this discussion (among other things) about future
possibilities of improvement is a sort of reward in itself: to see
people react, contribute and exchange ideas about something you created.
I don't see an ambiguous comment here and there as reason enough for you
to frown. :-)

Regards,
Tomislav