[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Syntax pattern
I added the following to the Introduction in the Syntax document,
under "Schema Patterns":
> The syntax follows certain repeating patterns:
>
> * "Entity" elements form the "top-level" elements, which can
> sometimes be contained in other entity elements. Examples of
> Atom entities would be <site>, <feed>, <entry>, <person>, and
> <archive> (complete set to be determined).
>
> * "Property" elements are direct children of their entity
> elements. Examples of property elements would be <title>,
> <summary>, <link>, <content>, etc.
>
> * Property and entity elements are all fully qualified, either
> in the core Atom namespace or an extension namespace. Examples
> usually show elements in the Atom namespace using a default
> namespace on the root entity element.
>
> * <content> and content-like elements (title, subtitle, and
> summary) have an associated a MIME type, mode of encoding,
> language, and either a value inline or through a URI
> reference. A <content> element may have a relation that
> indicates whether the content is an excerpt, preview, thumbnail,
> or otherwise not the entire content.
I thought it would be a good inaugural message to the list, to start
with a general discussion as to the style or pattern to the schema
rather than jumping straight to individual element-content models.
As always, these are not cast in stone. There is still discussion in
EscapedHtmlContent about what <content> may hold (is it limited to
just *ML, or should *ML types be special-cased). There is discussion
in ContentProblems regarding the distinction between whole and partial
entities. ContentDiscussion appears to have a simple majority
preferring only one <content>, but current examples are
multiple-contents with an undefined model for selecting a
representation. The utility of <content src='...'> appears to be in
question, but if it is useful, is a MIME multipart wrapper not far
behind (a la SOAP with Attachments)?
Since the decision was made to specify model and syntax in one
document (SyntaxConsiderations), I presume this list covers both?
-- Ken