Il 23/12/2009 11:28, Piero Sartini ha scritto:
>> I have an idea, may be a controversial one. I've seen much about Tapestry
>> and think it's a nice Framework. But what I still don't understand is why
>> it's not widely in use. It existed before Wicket but Wicket is far more
>> popular. I don't think it's only because it's users are very vocal. There is
>> also a saying that a good product sometimes sells itself.
> 
> I've also thought about this, and my only conclusion is that Tapestry is too
> difficult to master. Especially when it comes to tapestry-ioc and
> getting productive
> with it.

As a newbie, I have to say that learning Tapestry (5) is actually a
little bit more complicated than what you could expect after having read
the available "marketing" documentation. Maybe this (apparently steep)
learning curve has kept the "masses" of developers/users away from Tapestry.

Nevertheless, this apparent difficulty is in large part just a matter of
documentation and/or examples. From this point of view, Wicket seems to
be much more appetizing. Just have a look at these pages:

http://wicketstuff.org/wicket13/
http://sourceforge.net/projects/wicket-stuff/files/

It is not surprising that many developers could prefer a framework that
supply them with a lot of working, ready-to-use components and/or a lot
of code examples.

BEWARE: I'm not saying that Wicket /is/ better or more complete than
Tapestry. I'm just saying that Wicket is /presented/ (or /offered/) to
the public in a better way.

I think that it is possible to fill this gap, for example using Tynamo
and AppFuse as examples/codebases. At the end, it is just a matter of
documentation and support.

> I am not talking about building the frontend and components.. this is
> easy and IMHO
> the thing where tapestry really shines.

I agree. Tapestry is much more elegant than many other frameworks from
this point of view. This part of Tapestry should not be under discussion.

I do not know if the IoC container is the real and sole source of the
scarce appreciation of Tapestry (if even exists such a scarce
appreciation) but... see below.

> But there are way too less integrations - see my tapestry-jpa experiment for 
> an
> example. It works, but it would need some love from some smarter people than 
> me
> to get proper unit tests and some missing parts.
> 
> If I look at Wicket or other frameworks there are lots and lots of
> integration modules
> for just everything. Why is that? My answer is because it is way
> easier to write them.

I'm afraid you are right: /this/ seems to be the main, real weak point
of Tapestry. I do not know Tapestry well enough to have a solid opinion
about this topic but it seems to me quite evident that writing an
integration module is somehow (much?) more difficult than it should be.

Unfortunately, this can be a heavy limit for Tapestry. A real webapp
very often uses many external modules and the simple fact of not having
a easy way to write and/or integrate them can be a good reason for
abandoning this framework.

I came from the Python world and, as you probably know, a large part of
the success of Python is related to the easyness of integrating existing
C and C++ library with the Python interpreter (using SIP or SWIG). It
seems to me a lesson to learn.

I wonder: is it possible to improve the existing integration mechanism
of Tapestry (that is: the existing IoC container)? How? Or should we
replace it with a different/new one? Which one?

JM2C

-- 

Alessandro Bottoni
Website: http://www.alessandrobottoni.it/

Who wants to remember that escape-x-alt-control-left shift-b puts you
into super-edit-debug-compile mode?
     -- (Discussion in comp.os.linux.misc on the intuitiveness of
commands, especially Emacs.)

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
For additional commands, e-mail: users-h...@tapestry.apache.org

Reply via email to