view release on metacpan or  search on metacpan

Design.md  view on Meta::CPAN

# Articulate - Design

This document may be a bit cryptic! It's not a spec and it's not documentation (in the sense of a published interface) but a notepad of design ideas and aspirations, not all of which may come to fruitition.

Some of it is more formal, other bits are more stream-of consciousness.

Since writing much of this, implementation has moved forwards and the docs and bugtracker may be better sources of info.

## Synopsis

Articulate is a Content Management Framework. It provides a service which forms the backbone of your content management solution. Wherever possible the 'moving parts' are replaceable using plugins. It is intended to increase the 'whipuptitude' of the...

Articulate is an expression of the idea that your content management solution is foremost an API:

- it should not place arbitrary restrictions on the front end
- it should not place arbitrary restrictions on role of the content (blog, issue tracker, wiki)
- it should not place arbitrary restrictions on the content-types you want to host, or how you want to edit them
- it should not force you into a url schema you don't want.
- It should be easy to add an Articulate component to a project with other functions.


## Qs

### How to convert a blob to HTML?

The blob must be of a type.

### Can a blob be edited?

Depends on the blob type. If it is a subclass of file, probably not.

If the blob type has an associated edit method - this needs to make the browser load some js and maybe pass some arguments. Eg. load the XML editor and configure it.

### Can a blob be validated?

The blob type needs an associated validator writing.

### Does metadata need validating?

Ultimately yes. It may need to conform to several progressively more restrictive schemas.

### Can we create structured/HTML-encoded results?

Each section has a type.

Groups could be on a zone basis, e.g. public/authors (roles?)

What about Groups of groups, e.g. "developer" across projects

## Technical Specification

### Implementation

Create a content interpreter which gets meta and content together, and does things like run the XSLT. For this we really need the content retrieval to be OO.

The intepreter should be configured with a list of converters.

It will take the content and determine if it can convert the contents into HTML using the tools it has. If it cannot, it will offer a file location for download.

### Components

How are components (sections, comments, etc.) stored, loaded, configured, etc?

Does this include metadata extractors?

Problem: If you load a section, you need to run all the interpreters

Before the response is passed to the template, the components are loaded in order.

$component->process( $response ); # the response is mutated in-place

Components also need to register routes. Do they do so in a separate package? How about a method which modifies the route map?

### Events

An event is when something is done. It is not a hook: there is no possibility of interrupting the event.