Re: [xml-dev] RE: James Clark: XML versus the Web

[Michael Kay <mike@saxonica.com>]

> Assume that the complexity of a technology can be (crudely) measured using the length of its specification--the longer the specification, the more functionality that must be implemented and therefore the greater the complexity.

Fair as a first-order estimate, but it's a very crude metric.

In XPath 1.0, the count() function was specified in one line.

In XPath 2.0, it was 10 lines.

In XPath 3.0, it's 17 lines.

The function hasn't become any more complex in the meantime; it's just 
specified and explained more carefully. In fact the function has become 
simpler - in XPath 1.0 there was an error condition (argument not a 
node-set) which the specification didn't bother to mention, and which no 
longer arises in 2.0 and 3.0.

But I don't quarrel with your general premise. Standards, like software, 
become bloated over time, and the time comes when they collapse under 
their own weight and get replaced with something simpler.

Michael Kay
Saxonica