[Maypole] TPF Wrap-up

[Simon Cozens]

As the observant of you will have noticed, it is now no longer April. I
received $1000 funding from the Perl Foundation which paid my expenses
for the month of April while I hacked on a bunch of Maypole things. Despite
having my work cut short by the dreaded RSI, I feel I managed to achieve
enough to justify the grant:

    * Over 20,000 words of a user manual
    * Five articles on Maypole case studies (2 perl.com, one in last month's
      TPJ, one to go out to TPJ this month, and one for IBM DeveloperWorks.)
    * An e-commerce application, and one portal application ported from
      ASP.NET

In addition, the code enhancements I've been able to make by working on
Maypole full time have led to much increased stability and brought Maypole
to enterprise-readiness, as evidenced by the first Maypole-related job being
posted on jobs.perl.org a few days ago.

Although my current funding has run out and Maypole is now going to return to
being a free-time project for me, there's a small possibility that I'm going
to receive some funding from a computer consultancy to continue Maypole's
development.

With that in mind, I'd like to think about some longer-term directions for
Maypole. There are short and medium term goals in the todo list, but I want to
set out my vision for how Maypole is going to develop from here.

First, I want to produce some more examples of model classes that aren't based
on databases at all; I will be writing an NNTP interface for perl.org and this
may end up being Maypole based.

There needs to be a way of inheriting from Maypole applications, and of
packaging Maypole "components"; requiring a login and authentication, or
adding a shopping cart to an application, should be a simple matter of
'use'ing the component and configuring appropriately.

This in turn requires Maypole to sort out where it puts factory components,
and how Maypole itself gets packaged for install.

Finally, I'm slightly unhappy about the namespace clash between actions and a
model class's methods; witness the ugliness in Maypole::Model::CDBI::delete.
Decoupling this with a has-a relationship is a long-term goal, possibly
looking ahead several major stable revisions.

It only remains for me to thank the Perl Foundation once again for allowing me
time to work on Maypole. I hope that their contribution has translated into
your benefit.

Simon

-- 
IDIOCY:
    Never Underestimate The Power Of Stupid People In Large Groups

                                                    http://www.despair.com