[PJUG Javamail] ORBlogs team seeking help

[Craig McClanahan]

On Tue, Oct 14, 2008 at 8:53 PM, John Metta <mettadore at positorio.us> wrote:
> Greetings,
>
> I'm John Metta, one of the people working to try to bring ORBlogs back
> to life. For those who don't know, ORBlogs is an Oregon Blog Aggregator
> created by Paul Bausch (http://orblogs.com). Paul couldn't keep managing
> the site and a number of us bloggers on Twitter and elsewhere took up
> the cause to re-create it.
>
> One issue we are having is not enough Java experience. There is an
> excellent Digg+aggregator called BigBark (http://bigbark.net) created
> entirely in Java by Patrick Lightbody <http://twitter.com/plightbo>. He
> can't take an active role right now, but has donated the code to the
> cause and is allowing us to open source it. We've created a project page
> for the code at http://code.google.com/p/orblogs/.
>
> We're looking for some experienced Java developers who have some time to
> help us get this code up and running, and to help things along as we
> come up to speed. We think there are people and experience enough to
> maintain the codebase with one or two lead developers who are strong in
> Java taking an active role.
>
> This is an #afterhours project for all of us, so we're looking for some
> other people who'd be willing to help out. Patrick doesn't think the
> code has too much work to be ready for prime time- then we need to hook
> up our database and merge parsers created by A.J
> <http://twitter.com/linuxaid>.
>
> Anyone with some time who's interested in helping us out, give me a shout.
>
> Thanks,
> -John Metta

John (and others),

OK, I'll bite ... I'm local (Oregon native, born in Salem and moved to
the Portland area when I was three), and have a "wee" bit of Java
experience :-), although more in creating frameworks (Struts, JSF)
than applications.  I'm also currently working on some cool cloud
computing related stuff for Sun (about which I can't provide any more
details at the moment).

This sounds like an interesting project ... but in thinking about it a
bit, I would find it a *lot* more interesting if the architectural
direction were changed a little bit, in two directions:

* This seems like a natural sort of application to split into two
pieces -- a back end
  with RESTful APIs for managing the data, and one or more front end
applications
  providing UI on top.  There might also be some interesting
opportunities to link this
  functionality in with some of the social networking sites and APIs
that are emerging.

* If this kind of application becomes popular, it is likely to quickly
outgrow a single
  webapp instance, so we should build in scalability (and redundancy)
from the get go.
  Trying to do this after the fact, in the midst of a crisis driven by
rapidly increasing
  use of a live system, is incredibly painful.

My personal interest is more on the back-end service implementation
stuff (along with the intricacies of architecting in scalability).  I
will also confess a bias towards technologies I have been using lately
for this kind of thing (Jersey[1], the implementation of JAX-RS[2]
that just went final with version 1.0, plus EJB3 session beans on
Glassfish).  Besides making it *really* easy to create RESTful web
services, Jersey also provides support for using JAXB-generated
JavaBean classes for the data being sent back and forth, which can be
used to create a nice Java-oriented client SDK (which can then be
consumed by the front end UI applications written in Java).

For the front end, I don't have strong opinions about a favorite
Javascript+Ajax framework, and would happily defer to others who have
stronger feelings here.  But I would like to see more than one UI ...
perhaps one based on Swing, or perhaps even one based on Ruby on Rails
(another technology I've enjoyed using over the past couple years).

As with everyone else, this would be an after hours thing for me as
well, but I've got a pretty good idea of how I would see putting the
back end together (but need to bone up some more on the data model and
business logic before trying to define the RESTful APIs).  I would
enjoy working with others on this, especially with folks with a
passion for the UI part (if you do it, I don't have to :-).

What do you think?

Craig McClanahan

[1] https://jersey.dev.java.net
[2] http://jcp.org/en/jsr/detail?id=311



> @Mettadore <http://twitter.com/Mettadore> on Twitter
> http://positorio.us
> _______________________________________________
> Web Site - http://www.pjug.org/
> Javamail mailing list
> Javamail at pjug.org
> http://www.pjug.org/mailman/listinfo/javamail
>