That looks okay although schemas are not my strong suit.
I do believe you need minOccurs="0" and maxOccurs="unbounded" on <xs:element 
name="risk" type="riskType" />
(If there must be one risk the you can omit the minOccurs attribute.)

-- 

Joe Fawcett (MVP - XML)

http://joe.fawcett.name


"Adam David Moss" <no@e...> wrote in message 
news:O$a1shv2IHA.2332@T......
> Ok,
>
> So working on the advice so far I've come up with:
>
> <?xml version="1.0" encoding="UTF-8" ?>
>
> <xs:schema elementFormDefault="qualified" 
> targetNamespace="http://www.anurl.com" version="1.0" 
> xmlns="http://www.anurl.com" xmlns:xs="http://www.w3.org/2001/XMLSchema">
>  <xs:element name="risk" type="risksType" />
>
>  <xs:complexType name="risksType">
>    <xs:sequence>
>      <xs:element name="risk" type="riskType" />
>    </xs:sequence>
>  </xs:complexType>
>
>  <xs:complexType name="riskType">
>    <xs:sequence>
>      <xs:element name="description" type="xs:string" />
>      <xs:element name="response" type="responseType" />
>    </xs:sequence>
>
>    <xs:attribute name="severity">
>      <xs:simpleType>
>        <xs:restriction base="xs:string">
>          <xs:enumeration value="Low" />
>          <xs:enumeration value="Moderate" />
>          <xs:enumeration value="Substantial" />
>          <xs:enumeration value="Severe" />
>          <xs:enumeration value="Critical" />
>        </xs:restriction>
>      </xs:simpleType>
>    </xs:attribute>
>
>    <xs:attribute name="effectiveDate" type="xs:date" />
>  </xs:complexType>
>
>  <xs:complexType name="responseType">
>    <xs:sequence>
>      <xs:element name="description" type="xs:string" />
>    </xs:sequence>
>
>    <xs:attribute name="severity">
>      <xs:simpleType>
>        <xs:restriction base="xs:string">
>          <xs:enumeration value="Normal" />
>          <xs:enumeration value="Heightened" />
>          <xs:enumeration value="Exceptional" />
>        </xs:restriction>
>      </xs:simpleType>
>    </xs:attribute>
>  </xs:complexType>
> </xs:schema>
>
> Is this the correct way to lay this out or have I missed something?
>
> Many thanks.
>
>
> "Peter Flynn" <peter.nosp@m...> wrote in message 
> news:6cqgadF3fal8pU1@m......
>> Joe Fawcett wrote:
>>> I just think attributes are better for meta-data and single values. The
>>> only reason that I'd put description as a text node is that long text
>>> values are unwieldy in attribute values.
>>
>> I'd agree with Joe (and see
>> http://xml.silmaril.ie/developers/attributes/). I also favour terseness,
>> no matter what the Spec suggests :-) I'm really not convinced of the
>> need to repeat the locus as a prefix on every element type name:
>> redundancy is all very well, but it makes the file very cluttered.
>>
>> <risks>
>> <risk severity="low" effective="yyyy-mm-dd">
>>   <description></description>
>>   <response severity="high">
>>     <description></description>
>>   </response>
>> <risk>
>> ...
>> </risks>
>>
>> This also means you only have to define description once.
>>
>>> It's not something I'd get into a fight over though :)
>>
>> No, but it's the kind of thing that leads to long sessions in the pub at
>> XML conferences :-)
>>
>> ///Peter
>>
>