Re: Alternative to x!y

Norman Walsh wrote:
> / Alex Milowski <alex@milowski.org> was heard to say:
> | Two axis names does not seem like too much work to me.  We'll have to
> | describe the semantics anyway.  If we have our own syntax, we'll
> | have to describe that grammar.  Seems like the amount of work would
> | be the same.
> 
> Would those additional axes be available in test expressions?
> 
> I imagine that many implementors (myself included) are hoping to use
> an off-the-shelf XPath implementation. That's not going to be possible
> if we add axes. Besides which, making the source attributes into XPath
> expressions opens up a whole can of worms with respect to arbitrary
> XPath expressions, dynamically generated step and port names, etc.,
> that I'd just as soon leave in the tin, preferably stuck on the back
> shelf of some dark closet that we never have to open. :-)

You'd have to build your own parser for x!y anyway... small issue to
make it do step::x/port::y.

It isn't XPath in as much as Schema component expressions aren't XPaths.
They just borrow a similar syntax and concept.

--Alex Milowski

Received on Wednesday, 23 August 2006 19:42:41 UTC