Screaming Penguin

The way I see the plug ins working for the CMS is like this:

An XML conf file in the WEB-INF directory that lays out an ordered list of something that extends a PluginProcessor interface. The interface would take a wrapper object with a DOM tree of the content and some configuration information -- the URL for the document, the user being rendered to, categories (more on that in a sec). It could then manipulate it and return it for passage to the next plug in.

This would happen on both "in" and "out".

Now, on the "in" path, I kind of see this:

1. Send an XML-RPC update ping to Technorati and Blogs.com and whatnot.

2. Add to a del.icio.us "general user" the story with the categories as tags.

3. Create a post in jForum with the story contents and persist a StoryID-JForum URL map somewhere.

On the "out" path:

1. Add the proper "technorati tags" links so it can classify the story properly.

2. If it is a "full view", grab the RSS for the JForum post and include the comments on the page.

3. If it is a "full view", grap the Trackback, Pingbacks and include them on the page.

4. Perhaps if we had set up JSPWiki somewhere in the future it could look for keywords and auto-link to Wiki entries. (Silly idea, but it gives you an idea of the flexibility).

5. If it is not a full view, include the comment/trackback counts.

Again, I don't have any real design sketches for this, and certainly no code, but this is the idea.
1. Add technorati