XML query requirements cover a broad range of areas. The position paper by Allaire focuses on issues related to interoperability and integration with other W3C initiatives and requirements for mass-market use of XQL.

XML query requirements cover a broad range of areas. This paper starts out by focusing on some issues related to the interoperability and integration of XQL with other W3C initiatives, in particular:

• XSL,
• XLink and XPointer,
• XML Namespaces, and
• XSchema.

The paper concludes with a brief discussion of some requirements related to the mass-market use of XQL.

The query syntax in XQL and the match pattern syntax of XSL should be the same. To our understanding, there seems to be a general consensus on this requirement by both the XSL WG and the vendors that jointly proposed XQL.

There seems to be some area of overlap between XQL and XLink/XPointer as far as the identification of chunks of XML is concerned. This area should be mined with at least two questions in mind:

• Can one (or should one be able to) link against the result of an XQL query in some more sophisticated fashion than the HTML <a href= "..."></a> example given in the XQL submission, and
• Can there be some reconciliation of the XQL query and XPointer addressing syntaxes (perhaps as a subset of XQL).

Namespace specifications combined with namespace defaulting can modify the meaning of XML syntax both within the bounds of the element hierarchy (when applied to element names) and across the element hierarchy (when applied to attribute names). XQL queries can return any part of an XML document including one where the XML syntax uses namespace defaulting. XQL must provide a mechanism for preserving and returning namespace information to the application making the query.

Note: XLink and XPointer need to address the same issue.

XQL queries can return parts of an XML document that, taken on their own, will not validate against the document schema. The XQL initiative should investigate the options for preserving and returning schema information. Preservation of datatype information is required for smooth integration with non-declarative processing models.

Note: XLink and XPointer need to address the same issue.

The XML-related W3C initiatives should use the Web as an example of what it takes to create a successful computing platform. Simplicity, pragmatism, and ubiquity characterize the Web. The XQL initiative should embrace these characteristics as requirements. In particular, this means that the XQL specification should be constrained enough to the point where there should be no barriers to the introduction of ubiquitous, very low cost implementations for querying XML documents and DOM sub-trees.

Another area of concern regarding the mass-market use of XQL is the exposure of data source schema information. In this context data source denotes anything that XQL can query and schema denotes some meta-information about the organization of all XML data in the data source. One of the barriers to building good tools for automatic/visual query generation in the relational database world was and still is the fact that database schema (catalog) information was difficult to extract in a standardized fashion. To achieve interoperability between XML data sources and the tools that manipulate them some standardized mechanism for schema extractions needs to exist. This is a complex requirement to address and it should probably be apportioned to an XQL version other than 1.0.