-----Original Message-----
From: Von Riegen,
Claus [mailto:claus.von.riegen@sap.com]
Sent: Mon 08-Dec-03 22:16
To: Rogers, Tony; uddi-spec@lists.oasis-open.org
Cc:
Subject: RE: [uddi-spec] Our current XML schema is now inconsistent
with the spec?
It is not clear to me that we made such a decision. Which
section of the specification states that the element order doesn't
matter?
Claus
-----Original Message-----
From:
Tony.Rogers@ca.com [mailto:Tony.Rogers@ca.com]
Sent:
Sonntag, 7. Dezember 2003 12:08
To:
uddi-spec@lists.oasis-open.org
Subject: [uddi-spec] Our current XML schema
is now inconsistent with the spec?
When we decided to state that we
would accept the contents of a structure (using the term loosely) in any
order, we did not update the XML schema to reflect this. Consequently, the
schema is now inconsistent with what we say in the specification.
The
schema uses xsd:sequence, and that specifies that the elements must occur in
the order they appear within the sequence specification.
I not
sufficiently familiar with XML schema to be certain that there is no simple
way around this issue, but I fear that this is the case. Especially if we wish
to specify that we are unconcerned with the order of the elements, but that
any occurrences of a particular element must be together.
For example,
if we consider businessInfo, which currently
contains:
<xsd:sequence>
<xsd:element ref="uddi:name"
maxOccurs="unbounded"/>
<xsd:element ref="uddi:description"
minOccurs="0" maxOccurs="unbounded"/>
<xsd:element
ref="uddi:serviceInfos" minOccurs="0"/>
</xsd:sequence>
we
note that it will only accept one or more names, followed by zero or more
descriptions, followed by zero or one serviceInfos.
If we follow what
is specified in the current version of the UDDI V3 Specification, however, it
should accept a businessInfo that contains a description followed by a name -
right?
The interesting part comes when we consider a businessInfo that
contains a description followed by a name, followed by another description -
do we want to accept or reject such a construct? It is not all that difficult
to construct a schema to do either (I think), but neither schema is
particularly easy to read.
I would assume that the canonical form of
the XML (as required, for example, for verification of a digital signature)
would be the form that is accepted by the current schema.
The more I
ponder this, the more I wonder if we should undo that particular change to the
specification, and thereby require that the elements of a structure be in a
specific order. It's less flexible, but it's simpler.
What do you
think?
To unsubscribe from this mailing list (and be removed from the
roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/uddi-spec/members/leave_workgroup.php.