Will XML Schema 1.1 solve the problem?

The paper starts with an assessment of XML Schema 1.0. What do people want from the technology, and is it meeting their requirements? Clearly it must be doing something useful, because it has been so widely adopted. It hasn’t suffered the fate of some standards of being ignored. But at the same time it’s widely criticized on two almost contradictory grounds: it’s an immensely complex specification, and yet at the same time it often fails to provide the basic features that users are looking for.

We’ll take a quick look at the original design objectives for XML Schema, but we’ll concentrate our attention on what people actually want from it today (the differences are instructive). There are basically two applications for XML Schema today: validation of instances, and definition of data types for XML processing, whether by use of data binding to languages such as C++ and Java, or as the type system that underpins XQuery 1.0 and XSLT 2.0. We’ll examine where the specification lets these user communities down in terms of not providing the features they need.

We’ll then provide a quick tour of the new features of the 1.1 specification, some of which are impressive: assertions, conditional type assignment, open content models, interleave structures, and more.

And then we’ll try to summarize by comparing the two lists, so that we can assess to what extent the new facilities will satisfy the punters. Which in turn will give us the opportunity to conclude with some requirements for the future.