Hi Ray,

you are right in your assumption - the XML Schema spec defines rules ( 
http://www.w3.org/TR/xmlschema-1/#Complex_Type_Definition_details, 
3.4.2. ) which do not allow any other interpretation: just as if the 
type were xs:string, any character content is valid, and no child 
elements are allowed.

In the example case you showed, where the type definition is local 
(anonymous), empty mixed amounts to an eccentric way of saying 
type="xs:string". However, if the type definition is global and has a 
name, such an empty mixed type makes sense, because it can be used as 
base type for type derivations which are not possible if the base type 
is xs:string.

Namely, the empty mixed is a complex type and can therefore be extended 
by adding child elements, and it can also be restricted to a simple type 
which is not derived from xs:string. Both things are not possible if the 
base type is xs:string, which is a simple type and can only be 
restricted to subtypes of xs:string, or extended to a type with 
xs:string content plus attributes.

With kind regards,
Hans-Juergen Rennau


*Date:*         Mon, 22 Feb 2010 11:19:38 -0500
*Reply-To:*     "SRU (Search and Retrieve Via URL) Implementors"
              <[log in to unmask] <http://listserv.loc.gov/cgi-bin/wa?LOGON=A2%3Dind1002%26L%3Dzng%26T%3D0%26P%3D1514>>
*Sender:*       "SRU (Search and Retrieve Via URL) Implementors"
              <[log in to unmask] <http://listserv.loc.gov/cgi-bin/wa?LOGON=A2%3Dind1002%26L%3Dzng%26T%3D0%26P%3D1514>>
*From:*         "Ray Denenberg, Library of Congress" <[log in to unmask] <http://listserv.loc.gov/cgi-bin/wa?LOGON=A2%3Dind1002%26L%3Dzng%26T%3D0%26P%3D1514>>
*Subject:*      Explain schema question
*Comments:* To: "SRU (Search and Retrieve Via URL) Implementors" <[log in to unmask] <http://listserv.loc.gov/cgi-bin/wa?LOGON=A2%3Dind1002%26L%3Dzng%26T%3D0%26P%3D1514>>
*Content-Type:* text/plain; format=flowed; charset="iso-8859-1";
              reply-type=original

In the OASIS committee working on SRU 2.0 we are in the process now of 
taking a critical look at Explain and we have a technical question about 
the schema. Perhaps one of you Explain implementors might have some 
insight. (There no longer seems to be anyone claiming ownership of 
Explain, and if nobody steps forward, I am going to claim ownership, on 
behalf of the SRU maintenance agency, within the next few weeks.) Our 
question is this There are several datatypes whose definitions look like 
this: <xs:element name='host'> <xs:complexType mixed='true'> 
</xs:complexType> </xs:element> What does it mean? How would it differ 
from: <xs:element name='host' type = "xs:string"> We know what 
mixed='true' means, if for example, you had <xs:element name='host'> 
<xs:complexType mixed='true'> <xs:sequence> <xs:element name="a"/> 
<xs:element name="b"/> </xs:sequence> </xs:complexType> </xs:element> 
Then you could have an instance of element <host> that would look like: 
<host> (text preceding element a) <a>(content of element a) </a> (text 
between elements a and b) <b>(content of element b)</b> (text following 
element b) </host> In other words, mixed content and elements, which 
would not be allowed without mixed="true" But what does the first mean? 
The complex definition does not define any elements, and it seems to us 
that it degenerates to nothing more than a string. Nothing else makes 
any sense. Anyone have any other interpretations? Thanks. --Ray