On 12/21/05, Michael Kay <mike@xxxxxxxxxxxx> wrote:
> > Just to be clear on this one, there is no such function as
> > op:subtract-dates() (or any prefixed with op:) and it's in the CR
> > purely to demonstrate $earlier and $later?  Is that what you are
> > saying?
>
> There is no function with this name available to users (and to make this
> perfectly clear, we refrained from telling anyone what the namespace uri
for
> "op:" was). It's a formal specification device to define the semantics of
> the "-" operator when the arguments are dates.
>
> To be honest, the way the specifications of XPath operators are described
> was for the convenience of the authors, not the readers. (Why does it take
> more than ten times as many pages to specify XPath 2.0 as it took for XPath
> 1.0? Answer: because it was done by a bigger committee). I tried to get it
> changed, but the change was considered too risky. The argument is always
> that users who can't follow the spec can always buy themselves a book.

Thanks, that's an interesting way of doing it I guess.  It certainly
keeps the barrier up between the casual reader and the spec :)

(nb. I *know* the spec isn't written for the casual reader I'm just
making a joke - listing functions that don't actually exist is a great
new form obfuscation...)