Response to NB2

From RIF
Jump to: navigation, search

Nick,

Hi, thanks for the feedback and actually checking the schema.

Nick Bassiliades wrote:
> Hi,
> 
> I would like to congratulate the WG for a thorough and careful work
> on defining a rule standard for the web. It is important that the
> current standard is based on previous attempts, such as RuleML, so
> that interoperability with existing systems/prototypes is maintained.
> 
> Mostly, I would like to comment on the XML Schema for RIF-FLD (this may
> be present on other RIF-related Schemata; I didn't check).
> 
> Specifically, in the FLDSkyline.xsd file there are two lines:
> 
>    <xs:include schemaLocation="FLDBaseline.xsd"/>
>    <xs:redefine schemaLocation="FLDBaseline.xsd">
> 
> which actually should not co-exist, because xs:redefine implies that an
> external schema file is included, but some of the definitions in there
> are redefined. As a consequence, the above XML Schema file does not
> validate in e.g. XMLSpy. Actually, only the redefine is needed.

Thanks for catching this bug. While it validated in XSV, there were warnings. We thus dropped the xs:include and merged the two xs:redefine's in v. 1.4 of FLDSkyline.xsd (http://www.w3.org/2005/rules/wiki/FLD#Skyline_Schema_Module).

> Furthermore, I would like to remark that currently the XML Schema
> definitions for various elements are not opted for modularity and
> extensibility. For example, the definition of the element "Implies"
> encapsulates its complex data type, not allowing for redefinition by an
> extension of FLD.
>
> For example, I'm working on a defeasible logic extension of RIF-FLD,
> where I would like to add some attributes to indicate the rule type
> (strict, defeasible, defeater). However, since the complexType of
> "Implies" is not named, I cannot redefine it, including an attribute list.
> 
> I believe that since RIF is supposed to be a family of languages,
> extending one another, this extensibility should be taken into account
> for the XML Schemata.

The current XML Schema started with two top-level modules (FLDBaseline.xsd and FLDSkyline.xsd). Further modularization, possibly including the fine-grained modularity you are proposing, is a topic of upcoming efforts.

Please acknowledge receipt of this email to <mailto:public-rif-comments@w3.org> (replying to this email should suffice). In your acknowledgment please let us know whether or not you are satisfied with the working group's response to your comment.

-The RIF WG