Michael,

Thanks for the help.  It gives me something to work with, and to talk to the
document originator.

I couldn't find an answer in the FAQs, so I posted my original question
here. I also have some formatting issues, but I'll get to those as soon as I
figure out why this doesn't work.

I will contact EditiX about the Schema problem.

Michael Gosselin

> -----Original Message-----
> From: Michael Kay [mailto:mike@xxxxxxxxxxxx] 
> Sent: Tuesday, April 04, 2006 6:29 PM
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: RE:  Help deciphering someone else's XML schema
> 
> > I'm looking at some XSLT and XSD code from another company. 
> 
> The fact that you have some XSLT code doesn't make this an 
> XSLT question or make it on-topic for this list... 
> > 
> > <rr:CReport xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
> > xmlns:rr="urn:Correlagen:RightReport"
> > xmlns:ss="urn:schemas-microsoft-com:office:spreadsheet"
> > xsi:schemaLocation="urn:Correlagen:RightReport">
> > 
> > I'm a little rusty (well, very rusty) with the XML, XSD and XSLT 
> > namespaces, and could use a little help deciphering that line.
> 
> This start tag contains three namespace declarations. The 
> "rr" namespace is needed because it's used in the element 
> name. The "xsi" namespace is needed because it's used in the 
> attribute name. The "ss" namespace isn't obviously needed on 
> this start tag, but it's probably used somewhere deeper in 
> the content, judging from the fact that there's an "ss:Type" 
> mentioned later on.
> 
> xsi:schemaLocation tells the system where to find a schema. 
> In this case, however, it's incorrect. It's supposed to 
> contain a pair of URIs: first a namespace URI 
> (urn:Correlagen:RightReport) and then a location URI that 
> says where the schema for that namespace is to be found. The 
> location URI is missing, which is an error. You should raise 
> this with the document originator. 
> > 
> > When I take that file and put it into EditiX, and create a Schema 
> > (.xsd) file, I get some errors with validation.
> 
> If you asked EditiX to create a schema, then I would expect 
> it to create a schema that the instance is valid against. You 
> seem to be suggesting that it has created a schema which 
> defines the ss:Type attribute to be an xs:NCName, when 
> actually there's an ss:Type attribute that isn't an 
> xs:NCName. Unless I've read your post incorrectly, this would 
> suggest a bug in the EditiX schema generator, which you should report.
> 
> Michael Kay
> http://www.saxonica.com/