RE: Sequences in XSLT 2.0..

> XSLT 2.0 has introduced a major paradigm shift from XSLT 1.0 by
> introducing the concept of sequences. Now everything is a sequence.
> Earlier everything was a node-set. I wonder what was the need of such
> a major change? What does sequences concept achieve that node-sets
> cannot? Why was it considered neccessary to introduce in XSLT 2.0?
> 
> I'll appreciate if somebody could explain the rationale 
> behind sequences..

Several reasons:

(a) XML Schema allows list-valued elements and attributes, so a schema-aware
processor needs to handle lists of atomic values.

(b) The main reason it's so hard in XSLT 1.0 to do "sum of price times
quantity" (or any other sum of a computed value) is that XSLT 1.0 doesn't
allow collections of numbers (or strings), only collections of nodes. This
also makes string handling much harder, for example you can only provide a
usable tokenize() function if you support sequences of strings.

(c) XSLT 1.0 allows you to process a collection of nodes in sorted order,
but it doesn't allow you to save a sorted sequence of nodes in a variable.

Michael Kay
http://www.saxonica.com/