I'd like not to lose sight of what the actual use case for floating elements is.

We have a restriction on unordered sequences that all children must be elements.

I suspect some similar restriction applies to ordered sequences with one or more floating children. I.e., that all the children (floating or not) must be elements, must have initiators, etc.

That would rule out the complex interactions you are illustrating.

If people really truly have tons of very unordered things in complex sets and subsets, then an array of choice is probably the way they need to model it.



On Wed, Aug 14, 2013 at 7:44 AM, Tim Kimber <KIMBERT@uk.ibm.com> wrote:
Good question.Here are a couple more to consider, along the same lines:

Q1: Suppose that the group looks like this:

<xs:sequence dfdl:sequenceKind="ordered">
 <xs:element name="floating1" dfdl:floating="yes" type="xs:string"/>
 <xs:sequence ref="someGlobalModelGroup">
 <xs:element name="floating2" dfdl:floating="yes" type="xs:string"/>
<xs:sequence>

This is effectively an unordered group ( all except one member has floating='yes') but it contains a sequence group. Do we need to prohibit this?

Q2: Suppose that the group looks like this:

<xs:sequence dfdl:sequenceKind="ordered">
 <xs:element name="floating1" dfdl:floating="yes" type="xs:string"/>
 <xs:element name="nonFloating" dfdl:floating="no" type="xs:string"
             minOccurs="5" maxOccurs="5"
             dfdl:occursCountKind="implicit"/>
 <xs:element name="floating2" dfdl:floating="yes" type="xs:string"/>
<xs:sequence>

Here, the 5 occurrences of nonFloating could be scattered all over the group's content, with floating elements in between. Is that something that we want to allow? It seems inconsistent with the decision to re: occursCountKind and sequenceKind='unordered'.

regards,

Tim Kimber, DFDL Team,
Hursley, UK
Internet:  kimbert@uk.ibm.com
Tel. 01962-816742  
Internal tel. 37246742



From:        Steve Hanson/UK/IBM@IBMGB
To:        dfdl-wg@ogf.org,
Date:        14/08/2013 12:15
Subject:        [DFDL-WG] Extra restrictions needed for dfdl:floating 'yes' ?
Sent by:        dfdl-wg-bounces@ogf.org



Section 14.4 of the spec says:

"An ordered sequence of n element children with either n or n-1 of those children with dfdl:floating='yes' is equivalent to an unordered sequence with the same n element children with dfdl:floating='no'."

Does this mean that the extra restrictions we recently added for unordered sequences must also apply to floating elements?

If so these would manifest themselves as:

It is a schema definition error if an element with dfdl:floating 'yes' is an optional element or an array element and its dfdl:occursCountKind property is not ‘parsed’.

It is a schema definition error if two or more elements with dfdl:floating 'yes' in the same group have the same name and the same namespace.

Regards

Steve Hanson
Architect, IBM Data Format Description Language (DFDL)
Co-Chair, OGF DFDL Working Group
IBM SWG, Hursley, UK
smh@uk.ibm.com
tel:+44-1962-815848
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--
 dfdl-wg mailing list
 dfdl-wg@ogf.org
 https://www.ogf.org/mailman/listinfo/dfdl-wg

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
--
  dfdl-wg mailing list
  dfdl-wg@ogf.org
  https://www.ogf.org/mailman/listinfo/dfdl-wg



--
Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com