Print

---

Yes, that's a known bug which came up a little while ago. It is fixed in
the drafts http://samantha.oucs.ox.ac.uk:8080/srw1-1/ (XML Spy regarded
an extension of all as legal but it turns out it isn't!).

I'm afraid it is still pending on me to review the drafts at
http://samantha.oucs.ox.ac.uk:8080/srw1-1/, write up the changes and get
Ray to update the ones on LoC (the last either weeks have been somewhat
hectic.)

Matthew

> -----Original Message-----
> From: Z39.50 Next-Generation Initiative [mailto:[log in to unmask]] 
> On Behalf Of Robert Sanderson
> Sent: Tuesday, July 13, 2004 12:01 AM
> To: [log in to unmask]
> Subject: Schema issues
> 
> Futher to my comment about not being ready for 
> standardisation, see below
> message forwarded from Tom Habing at UIUC.
> 
> (Summary:  MSXML says our schema is fubar, and another 
> validator does the
> wrong thing)
> 
> Matthew, any comments?
> 
> Rob
> 
> ---------- Forwarded message ----------
> Date: Mon, 12 Jul 2004 11:18:42 -0500
> From: Thomas G. Habing <[log in to unmask]>
> To: Robert Sanderson <[log in to unmask]>
> Subject: Re: OAI-PMH registry and ZeeRex
> 
> Hi Rob,
> 
> Sorry its taken me so long to get around to this.  However, 
> in looking at this
> fix I seem to have discovered a potential problem with SRW XML Schema:
> 
> http://www.loc.gov/z3950/agency/zing/srw/srw-types.xsd
> 
> Being a good XMLer :-) I always try to validate my files 
> against the relevant
> schemas.  My tools of choice are usually the Microsoft XML 
> toolkit (MSXML 4.0)
> and the XSV validator from 
> http://www.ltg.ed.ac.uk/~ht/xsv-status.html.
> 
> The MS toolkit complains that the XML Schema is not valid, 
> and XSV does not
> give the expected result.
> 
> The XML Schema fragments relevant to your explainResponse are these:
> 
>    <xsd:complexType name="responseType">
>      <xsd:all>
>        <xsd:element ref="version"/>
>        <xsd:element ref="diagnostics" minOccurs="0"/>
>        <xsd:element ref="extraResponseData" minOccurs="0"/>
>      </xsd:all>
>    </xsd:complexType>
> 
>    <xsd:complexType name="explainResponseType">
>      <xsd:complexContent>
>        <xsd:extension base="responseType">
>          <xsd:all>
>            <xsd:element ref="record"/>
>            <xsd:element ref="echoedExplainRequest" minOccurs="0"/>
>          </xsd:all>
>        </xsd:extension>
>      </xsd:complexContent>
>    </xsd:complexType>
> 
> Assuming that these are valid (see the following), to me this 
> indicates that an
> <explainResponse> must contain at least a <version> and a 
> <record>, along with
> some optional stuff including <diagnostics>.  Therefore, I 
> was surprised when
> the XSV validator didn't complain about your original 
> suggested correction
> above, because it didn't contain either a version or record 
> element.  Then I
> tried validating it using the MS XML toolkit which complained 
> that the XML
> Schema itself was not valid.  When I looked up the error 
> message returned by MS
> XML, I got this web page
> http://support.microsoft.com/default.aspx?scid=kb;EN-US;q316635 which
> essentially explains that using <xsd:all> inside of an 
> <xsd:extension> is not
> legal.
> 
> Presumably, the developers of this schema have done some 
> testing with various
> validators, so I would be willing to accept that Microsoft's 
> interpretation of
> the schema spec is wrong, but it is also curious that the XSV 
> validator, while
> it doesn't complain of an error is unable to correctly 
> validate against the
> schema.
> 
> My suggestions would be to avoid using extensions of all 
> groups, such that the
> explain response becomes:
> 
> 
>    <xsd:complexType name="explainResponseType">
>      <xsd:all>
>        <xsd:element ref="version"/>
>        <xsd:element ref="diagnostics" minOccurs="0"/>
>        <xsd:element ref="extraResponseData" minOccurs="0"/>
>        <xsd:element ref="record" minOccurs="0"/>
>        <xsd:element ref="echoedExplainRequest" minOccurs="0"/>
>      </xsd:all>
>    </xsd:complexType>
> 
> Assuming that the record really should be an optional 
> element.  The other "all
> extensions" could be modified similarly.
> 
> In any case, I've made your suggested correction, ignoring 
> the fact that I
> can't get it to validate with the current schema.
> 
> Kind regards,
>         Tom
> 
>