Re: Welcome

[Greg Stein <gstein@xxxxxxxxxx>]

On Wed, Sep 08, 2004 at 11:54:19AM -0400, Joe Gregorio wrote:
>...
> What that means is that the output of this team is a series
> of Paces, all of which need to be accepted by the
> whole WG, which means discussion on [atom-syntax].
>...
> As a starting point for the discussion I would
> like to point everyone to the TypePad
> implementation of the Atom Publishing
> Protocol[3]. There is a large chunk of functionality
>...
> Categories. We also need to consider
> which elements are mandatory versus optional,
>...
> In addition from our own charter we have the
> following functionality we must address:
> 
>     * multiple authors for a feed
>     * multiple subjects or categories in a feed
>     * user authentication
>     * adding, editing, and deleting users
>     * setting and getting user preferences
>     * creating, getting and setting related resources such as
>         comments, templates, etc.

This is certainly a valid approach to begin discussion -- what is being
done today is a good demonstration of "what is needed?" However, I'd like
to propose a different path for tackling the problem. Something a bit more
methodical:

* identify and design the data model
  - what objects are represented?
  - what properties do they have?
  - what are the relationships?

* identify the operations needed to operate on the data model

* define the protocol around the above items

I think the last step will just kind of "fall out" once we have the first
two items. Once the operations and data is known, then it becomes
relatively trivial to express that within different protocol models (REST,
SOAP, etc).

This approach also defers the question of protocol representation and
allows us to concentrate on the fundamentals of what we're trying to do.

Any thoughts on trying this approach?

Cheers,
-g