On Wed, 15 May 2013 17:30:17 -0300, Lenny Primak <lpri...@hope.nyc.ny.us>
wrote:
using the CDI spec. I am not even sure that what you are calling
Distributed Configuration is even needed for the Web Framework.
Just take a look at Tapestry (the web framework) itself. Look at how a new
coercion, request handler, validator, bean block override, etc is added.
It's quite easy to do it and it involves no source changes, no
configuration files. Now imagine that before distributed configuration.
The only workaround I can think of is, for every *single* service that
wants some kind of distributed configuration, to create another one so
anyone who wants to contribute would inject this configuration service and
add stuff to it. For each service that needs distributed configuration,
we'd need another. More code. Lost of it. Argh. Distributed configuration
is a huge code saver when you want to have flexible code.
For background on why Tapestry-IoC was created, read
http://tapestry.apache.org/ioc.html, specially the "Why not X?" sections.
If you are implying that this is all so important, why isn't every
project on the planet using Tapestry-IOC?
Few people know it, unfortunately. And fewer people knew that it could be
used without Tapestry-the-web-framework. Of course, the name doesn't help
here. In addition, many of the best options for problem X aren't used
because some other option is not as good but it's more popular. Damn,
until few years ago people still used Struts 1, which is awful, even when
other way better Java frameworks could be used for production sites. A
similar question: why isn't everybody using Mac OS or Linux instead of
Windows?
I would be very happy using the Web Framework without Tapestry-IOC,
using just plain beans for configuration,
or even using CDI events to gather configuration.
I didn't know about CDI events. I'll read about it. Thanks for the tip. :)
Given the history of Tapestry, CDI wan't there and popular enough at the
begging to be used for this, but now it is.
CDI as a separate entity from Seam, which is/was tied to JSR, barely
existed at that time. It was first proposed in May 23rd, 2006 with the
first approved version in December 10th, 2009. Tapestry 5, including
Tapestry-IoC, was started around
I also know that no way in hell would HLS give up Tapestry-IOC, but it
is indeed , in this day and age,
is a barrier to entry to learning Tapestry (Web Framework)
As I said in another thread, you're suggesting replacing Tapestry-IoC with
CDI. If that was done, people would still learn one IoC framework in order
to learn Tapestry. CDI has a broader reach (in termos of concepts and
features) than T-IoC. Not much people use CDI now (I may be wrong, of
course). Given all that, I don't think replacing Tapestry-IoC with CDI in
Tapestry would turn Tapestry much easier to learn, if at all. And you'd
need to rewrite a lot of Tapestry code, which would need to get bigger. I
don't think that's worth the effort at all.
--
Thiago H. de Paula Figueiredo
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
For additional commands, e-mail: users-h...@tapestry.apache.org
