W3C home > Mailing lists > Public > public-prov-wg@w3.org > September 2012

Re: PROV-ISSUE-547: prov:records requires a specific order of elements [XML Serialization]

From: Hua, Hook (388C) <hook.hua@jpl.nasa.gov>
Date: Thu, 13 Sep 2012 08:11:26 +0000
To: Provenance Working Group <public-prov-wg@w3.org>
Message-ID: <CC76D01C.153A0%hook.hua@jpl.nasa.gov>
This suggestion would also make prov:records be more consistent with the
xs:choice used in the complex type Dependencies.

One suggestion would be something like:

<xs:complexType name="Records">
    <xs:choice minOccurs="0" maxOccurs="unbounded">
      <xs:element name="dependencies" type="prov:Dependencies"/>
      <xs:element ref="prov:activity"/>
      <xs:element ref="prov:entity"/>
      <xs:element ref="prov:agent"/>

Note that in the old prov-xml,
  <xs:element name="dependencies" type="prov:Dependencies" minOccurs="0" />
Thereby implying the default of maxOccurs="1".

But using an unbounded xs:choice here would permit the <dependencies>
element to be unbounded. That would not make sense. Unless we apply
flattening per PROV-ISSUE-551.


On 9/10/12 7:53 AM, "Provenance Working Group Issue Tracker"
<sysbot+tracker@w3.org> wrote:

>PROV-ISSUE-547: prov:records requires a specific order of elements [XML
>Raised by: Curt Tilmes
>On product: XML Serialization
>Right now, the schema records element:
>requires a specific sequence of elements, in order:
>activities, entities, agents, dependencies.
>It would be less restrictive to allow, for example, entities to be
>specified prior to activities (as in some of the examples).
>I propose to wrap the "activity", "entity" and "agent" elements with an
>xs:choice allowing any number of any of them to be specified in any order.
>(I'm actually fine with requiring dependencies to follow those elements).
>We can explain this ordering of elements clearly in the verbiage in the
>tech note.
Received on Thursday, 13 September 2012 08:15:08 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:18 UTC