On 11-12-09 12:36 PM, Nathan Bubna wrote:
On Fri, Dec 9, 2011 at 9:18 AM, Antonio Petrelli
<[email protected]> wrote:
Hello
2011/12/9 Greg Reddin<[email protected]>
I guess we're essentially trying
to reconstitute the project. It's basically like going through the
Incubator with Antonio and I (and Nathan) as mentors. Once the "new"
PMC is established and running smoothly we will likely elect a new
chair and Antonio and I, and I presume Nathan, will go emeritus and
let you guys rock on. In fact I will probably include a statement to
this effect in our board report which is due this month.
Thank you Greg for your statement of confidence :)
I won't feed the discussion on leaving apache, especially since I've not
read any reason to leave so far. If there are, I want to know about
them! Please tell! (off-list if you wish)
I'm more interested in that part:
> But again, i insist that the choice must be made by
those who have vision for Tiles' future.
Indeed I believe that vision is what makes it a project, not lines of
codes or even a group of people working together (that would be a team).
I don't have a clear vision of the future either, just a number of
opinions, but I can share them in the hope of helping in forming that
vision.
First of all, Tiles has a well established position as a view
composition framework in the now traditional MVC2 pattern for java web
applications. It works smoothly with major MVC frameworks, including
Struts (its parent project), and Spring (including all latest versions),
even JSF to some extent (but JSF being a composition framework itself
favors other technologies), and other less used frameworks. It also
supports major templating languages: JSP, Freemarker and Velocity.
This should go on in the future, and be expanded as needed. The question
of which project is best suited to build and maintain the interfaces is
a tough one; historically the integration with MVC frameworks were
maintained by the framework's project, while the integration with the
templates were maintained by tiles.
In version 3 Antonio (thanks again) has layed out the basis of a request
abstraction to help with the integration of view technologies in various
environments. Servlet and Portlet where the initial targets, but we can
easily widen the scope. Besides my personal project (which involves
rendering the same tiles definitions both online through servlets and
offline by a background thread), I can imagine tiles applicable in any
situation where software is producing HTML. Just within Apache I can
think of a Camel Translator to create HTML-based emails, or the obvious
maven-site-plugin.
Another axis for development is the configuration. By moving from a
properties-based configuration to a java-based one, Tiles has opened
many options, but that power can be properly harnessed only by using an
IoC container. I'm using Spring with some success, and I'll try and
contribute that back in time, and I believe we should offer some support
for Guice and probably CDI as well, though I've not used those a lot.
These are my major ideas. Of course there are probably some minor
improvements/bugfixes that would come along the way.
Hope this helps,
Nick
