=head1 NAME

ROADMAP - The "SearchAll" Firefox extension project roadmap

=head1 DESCRIPTION

This document describes the current status and development plans of
the "SearchAll" Firefox extension project.

=head1 OBJECTIVES

"SearchAll" is a client-side site-by-site searching tools designed
as a Firefox extension.

As a side effect, we'll also develop a
general-purpose XUL application framework based on L<Template::Declare>,
L<jquery.js>, and L<Object::Declare>.

=head2 Project Overview

=over

=item *

Phase 1 - "Raw" view

This phase is finished.

=item *

Phase 2 - Architecture refactoring

This phase is finished.

=item *

Phase 3 - "Formatted" view

This phase is finished.

=item *

Phase 4 - "Merged" view

This is the current phase.

Phase 5 - "XX" view

Not started yet.

=back

=head1 Phase 1: 'Raw' view

The goal for this phase is to display the raw HTML pages
in 3 subwindows (C<< <iframe> >> or C<< <browser> >>) aligned in the same
line for the search engine sites being searched and
compared (i.e. yahoo.cn, baidu.cn, and google.cn).

=head2 Dependencies

=over

=item *

Use the C<onload> event for browser.contentWindow.

=item *

Use C<nsIWebProgressListener> for progressbar.

=item *

Solve the next-page problem in C<browser> and C<iframe>.

=back

=head2 Status

Done.

=head1 Phase 2: Architecture refactoring

The goal for this phase is to refactor the whole architecture.
The plan is to develop a C<XUL::App::Schema> module which will
be used in Makefile.PL and generate our own Makefile fragment.

=head2 Dependencies

=over

=item *

Implement a good SearchAll::Test and SearchAll::Logger.

=back

=head2 Status

Done.

=head1 Phase 3: 'Formatted' view

The goal for this phase is to add a new view (as a tab) called 'Formatted'.
In this tab, we will display the search results
returned from those various search engine sites in a uniformed
format instead of those sites' own HTML.

The list for each site is still displayed in a separate subwindow.

=head2 Dependencies

=head2 Status

Done.

=head1 Phase 4: 'Merged' view

The goal for this phase is to merge search results from various
sites in a single list and displayed in our own format.

=head2 Status

Not started.

=head1 RANDOM THOUGHTS

Below is a list of random thoughts appeared in the process of
development. They're not directly related to the objective of
this very project but they'll be benefitial in general.

=head2 DOM lib for SpiderMonkey

Develop a HTML/XUL/XML DOM library for the SpiderMonkey engine
using Perl. One potential way is

   HTML/XUL/XML source => Perl's HTML::TreeBuilder/XML::DOM
                       => JS code to build the DOM architecture
                       => SpiderMonkey.pm

=head3 Potential applications

=over

=item *

HTML/XUL/XML source postprocessor using JS (for instance, we can
use jquery.js here).

=item *

Test our JS code which manipulates DOMs.

=item *

Test our HTML/XUL/XML source's sanity. (It's particularly useful if
the HTML/XUL/XML stuff is got from a generator like L<Template::Declare>.)

=back

=head2 JavaScript language extention

A compiler which accepts a superset of JS and emits a fast
and portable subset. It's like something that jianingy++ mentioned before.

=head2 C<XUL::App>

A complete XUL application framework named C<XUL::App>. Unlike L<Jifty>,
it'll be mostly compiling-oriented so the runtime cost will be minimal.

=back