Document Root

A small comment regarding section 3.2, which may illustrate a much larger issue for schema as well as the XML effort in general.

In 3.2, the document states:

NOTE: We have not so far seen any need to reconstruct the XML 1.0 notion of root. For the connection from document instances to schemas, see Schema Validity.

For validating particular document instances this is certainly true.  However, this is a very useful piece of information if one is constructing tools to define document processing  based on the schema (for instance, a GUI appbuilder tool defining what you are to do with expected XML data to be seen by your application).  You would like to know which (probably) small subset of the defined elements may appear at the root level of the document, so as to provide appropriate navigational displays without (possibly) hundreds of element types which will not appear at that level.

The larger issue is that I am concerned with XML as a data interchange and with building programmer tools to manipulate it as such.  I am concerned that not enough attention is being paid to "toolability" in addition to ability to express the desired forms of structured data, transmit it, and validate it on the API level.  Certainly, the spec is open-ended enough to add extra information.  But I think the "base level" stuff ought to able to allow the construction of basic processing tools which would be able to operate on "vanilla" schema definitions not necessarily supporting a particular list of extensions, much like you can currently construct a reasonable GUI tool to operate on a vanilla relational DB through the table/column information available through an ODBC connection.

Received on Monday, 19 July 1999 17:57:02 UTC