Fix misleading ParticipantRef documentation

[hfjelstad]

Problem

The ParticipantRef element in the publication schemas is documented as
"Identifier of system requesting Data."

That wording only fits the request case. In a publication delivery the field
identifies the party that produced the data, not a requestor — so the
mandatory (delivery) use is documented incorrectly.

Additionally, the value is a code (NMTOKEN via siri:ParticipantCodeType),
not a reference to an object. There is no Participant type and no keyref —
the value is the identity. The ...Ref name suggests an object reference that
does not exist.

Fix

Short term (documentation only, non-breaking): update the xsd:documentation
to cover both publication and request:

External code identifying the source of the data being exchanged.

Applied to all four top-level publication schemas:

• xsd/NeTEx_publication.xsd
• xsd/NeTEx_publication_timetable.xsd
• xsd/NeTEx_publication-NoConstraint.xsd
• xsd/wsdl/NeTEx_publication-NoConstraint.xsd

The element name and type are left unchanged, so there is no schema break.

[skinkie]

Maybe do a pull request on the SIRI repo too?

[hfjelstad]

@skinkie SIRI does not seem to have the same issue:

ParticipantCodeType (NMTOKEN) | «Type for Unique identifier of participant.»
ParticipantRefStructure (extends code) | «Reference to Unique identifier of participant.»

So it could be changed to:

ParticipantRefStructure (extends code) | «External code identifying a participant.» ??

The base branch was changed.

[trurlurl]

To me, the following would make things even more clear: "External code identifying the sender of the message."

[hfjelstad]

@trurlurl i do agree in the context of NeTEx, but since it is used in SIRI as well it would be too narrow. ParticipantRefStructure in SIRI is used as base for ConsumerRef, SubscriberRef and ResponderRef,