I like the idea of specifying a type for the file, instead of a distinguished top-level element.

From: dfdl-wg-bounces@ogf.org [mailto:dfdl-wg-bounces@ogf.org] On Behalf Of Steve Hanson
Sent: Wednesday, May 06, 2009 1:22 PM
To: dfdl-wg@ogf.org
Subject: [DFDL-WG] DFDL: Applying DFDL annotations to elements

To apply DFDL annotations to a top-level element in a DFDL xsd, most modellers would use the dfdl:element dfdl:ref property to refer to a named dfdl:defineFormat block that set up the necessary defaults for all the DFDL properties. To avoid having to re-state the dfdl:ref property on every object that comprises the format, most modellers would also use the dfdl:complexType dfdl:ref property to scope the same dfdl:defineFormat block. The xsd would look like below.

<xs:schema ...>
<xs:annotation><xs:appinfo source=”http://www.ogf.org/dfdl/”>
<dfdl:defineFormat name=”textFormat1">
<dfdl:format encoding="utf-8" separator="\n" representation=”text” lengthKind="delimited" />
</dfdl:defineFormat>
</xs:appinfo></xs:annotation>
...
<xs:element name="textDoc" dfdl:ref="textFormat1" dfdl:lengthKind="implicit">
<xs:complexType dfdl:ref="textFormat1">
<xs:sequence>
...
</xs:sequence>
</xs:complexType>
</xs:element>
...
</xs:schema>

It's not possible to put DFDL defaults in scope for the whole format with a single dfdl:ref property. I think this is a side-effect of removing the dfdl:appliesTo property.

If this is thought to be an issue, there are a couple of options:

One is to say that a complex type can be the top-level object. This is the case with several XML based systems. It works with XML because the XML instance document provides the name of top level element in the infoset via its tag. This is not the case with DFDL where the name is commonly not carried with the format. So we'd have no name for the infoset.

Another is to provide a new property on dfdl:defineFormat, which says this dfdl:defineFormat is the default for all top-level objects in the xsd. Any top-level object that remained silent as to its dfdl:ref would get the default applied. I'm not sure whether this makes the model too opaque though. No more so than the existing scoping rules, I suspect.

Any other opinions or suggestions welcome.

Regards

Steve Hanson
Programming Model Architect
WebSphere Message Brokers
Hursley, UK
Internet: smh@uk.ibm.com
Phone (+44)/(0) 1962-815848

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU