Submitted for consideration by the W3C XML Query Language Workshop, 1998.
Agranat Systems produces web technology for embedded systems. We provide products that enable our customers products to act as HTTP/1.1 servers. At this time, the principle utility of this is to create network user interfaces intended for direct use through Web browsers.
With the advent of XML, there is increasing demand for integration between applications and web enabled devices. It is our hope that the XML Schema work will form the basis for unambiguous expression of data/document instances within a well defined framework that includes richer description of the content semantics than base XML provides.
Our interest is in standardizing a mechanism for expressing a query to produce an XML document, whether or not the target of the query (the thing being queried) is itself an actual XML instance. Ideally, the query itself would be an XML document.
It should be possible to determine what valid queries can be constructed by examining the set of XML schemas supported by a given potential query target (perhaps as supplemented by an XML document that describes or qualifies capabilities of the target).
It should not be a goal of an XML query language to govern formatting of the query result, other than by filtering the result through one or more associated style sheets; the style sheet mechanism should not be a component of the query mechanism. Ideally, a good boundary should be created between the CSS/XSL mechanisms and any new query language, perhaps separating (and removing) the query capabilities from existing efforts in order to provide for a clean division.