Print

---

I had somewhere the same problem, i.e. the (painful :-) dilemma:

a)	either fine control of the number of occurrences (obtained via 
<sequence>);
b)	or unconstrained order (obtained via <choice maxOccurs="unbounded">).

Maybe we should suggest W3C (for the XSD 1.1) the "natural" extension, 
something like:

<choice minOccurs="x" maxOccurs="y">
	<element name="e(p)" minOccurs="n(p)" maxOccurs="m(p)">... </element>
	...
	<element name="e(q)" minOccurs="n(q)" maxOccurs="m(q)">... </element>
</choice>

where the elements are allowed in any order, but WITH a constrained number of 
occurrences.

In this case, x and y are imposed (i.e. they could disappear ?):

x = sum (n(i), i = p..q)
y = sum (m(j), j = p.. q) 


What do you think ?

Dan

-----------------------------------------------------------------------
Dan Matei, director
CIMEC - Institutul de Memorie Culturala [Institute for Cultural Memory]
Piata Presei Libere nr. 1, CP 33-90
013701 Bucuresti [Bucharest], Romania
tel. (+40-21) 317 90 72; fax (+40-21) 317 90 64
www.cimec.ro


-----Original Message-----
From: Joe Altimus <[log in to unmask]>
To: [log in to unmask]
Date: Fri, 8 Aug 2008 14:53:00 -0700
Subject: Re: [MODS] schema use of xsd:sequence to define children of the 
cartographics element

> OK, I think I get it. There is no good reason for rigidly ordering these
> elements, we've just run into a limitation of the XML Schema standard.
> 
> On Fri, Aug 1, 2008 at 6:55 AM, Ray Denenberg, Library of Congress <
> [log in to unmask]> wrote:
> 
> > <cartographics> is defined as xsd:sequence because <scale> and
> >  <projection>
> > are not repeatable, but <coordinates>  is repeatable.
> >
> > <all> requires that each element occur zero or one time and would not allow
> > multiple <coordinates> elements.
> >
> > <choice maxOccurs="unbounded"> (which MODS uses in cases without such
> > constraints)  would not constrain <scale> and <projection> to maximum one
> > occurence.
> >
> > (Contrast this for example with <subject> ........
> >
> > <choice maxOccurs="unbounded">
> > <element name="topic"/>
> > <element name="geographic"/>
> > <element name="temporal"/>
> > <element name="titleInfo"/>
> >
> > etc.
> >
> > </choice>
> >
> > etc.
> >
> > .....any of these subject categories may occur any number of times,
> > including zero, and the whole set may occur in any order.)
> >
> >
> > ----- Original Message -----
> > From: "Joe Altimus" <[log in to unmask]>
> > To: <[log in to unmask]>
> > Sent: Thursday, July 31, 2008 5:30 PM
> > Subject: [MODS] schema use of xsd:sequence to define children of the
> > cartographics element
> >
> >
> > I notice that the MODS schema uses xsd:sequence to define the children
> > of the cartographics element:
> >
> >        <xsd:sequence>
> >               <xsd:element name="scale" minOccurs="0" type="xsd:string"/>
> >               <xsd:element name="projection" minOccurs="0"
> > type="xsd:string"/>
> >               <xsd:element name="coordinates" minOccurs="0"
> > maxOccurs="unbounded" type="xsd:string"/>
> >        </xsd:sequence>
> >
> > However, an example of <cartographics> in the MODS User Guidelines
> > does not conform to the schema:
> >
> > <subject>
> >  <cartographics>
> >    <coordinates>E 72°--E 148°/N 13°--N 18°</coordinates>
> >    <scale>1:22,000,000</scale>
> >    <projection>Conic proj</projection>
> >  </cartographics>
> > </subject>
> >
> > so I wonder if the example is wrong or the Schema (xsd:all instead of
> > xsd:sequence)? Is there a strong reason why the children should occur
> > in a rigid order in this case?
> >
> > Joe Altimus
> > Arizona State University Libraries
> >