XML Infrastructure Technologies | XML, Web Services, and the Data Revolution


Team-Fly

	XML, Web Services, and the Data Revolution By Frank P. Coyle
	Table of Contents

	Chapter 2. The XML Technology Family

InfoSet

InfoSet is the basis for a variety of XML technologies.

The InfoSet specification results from the fact that we're in a software world where it is important to be able to combine Web technologies for previously unanticipated purposes. This means specifications don't stand in isolation. W3C specs generally do one of two things: they describe what they do in natural language, most commonly English, or they provide a syntax or a grammar. Both approaches can prove difficult when trying to determine the exact meaning of XML specifications.

For example, consider the concept of empty elements in XML. The XML 1.0 specification says that one may either provide a start and end tag with no content, such as

 <foo></foo>

or use an empty element tag, such as

 <foo/>

In most XML vocabularies the distinction is irrelevant. However, there are certain situations where the distinction matters. For ex ample, some applications might want to treat <foo></foo> as signifying "data with the value NULL" and <foo/> as signifying "no data." The problem is, one cannot predict how XML will be used and what environment will be processing the XML.

InfoSet supports distinctions that are important when different XML specifications must interoperate .

While in one sense the two forms are equivalent, they are still different constructs. Is one interchangeable with the other? Most would say yes, they are interchangeable. However, it's possible to give one a semantic interpretation that makes them quite different. For example, it could be argued that the single-tag empty element emphasizes a lack of content, since it is impossible to provide content with just a single tag. The contentless start and end tags, on the other hand, are at least open to the possibility of content.

Secure communication requires unambiguous assumptions.

While it may seem that we're splitting hairs here, the problem is real. Consider, for example, a situation with possible legal implications. If I generate a digital signature to sign an XML document, do I want to invalidate that signature by changing <foo></foo> to <foo/> or vice versa? To answer such questions we need some way to talk about the structure of an XML document that goes beyond the grammatical rules of the language. We need to talk about the language in some universally communicable way. We need a metadiscourse. We need an InfoSet: a consistent set of definitions for use in other specifications that need to refer to the information in a well- formed XML document. (See the resources section at the end of this chapter for links to more detail on InfoSet.)


Team-Fly

Top