- From: Alex Milowski <alex@milowski.org>
- Date: Mon, 02 Oct 2006 10:27:53 -0700
- To: public-xml-processing-model-wg@w3.org
Erik Bruchez wrote:
>
> Alex Milowski wrote:
>
> > 1. Step must be able to refer to other steps that are
> > siblings (preceding and following) otherwise you
> > can't connected steps at all.
>
> "Preceding siblings" would be enough IMO.
And yet another example of where you wouldn't want this:
I have a pipeline with steps:
validate -> xinclude -> transform
but the input document doesn't validate. So I quickly
change, as an experiment, the input mappings to
xinclude -> validate -> transform.
Oh, but that pipeline won't compile because the "preceding sibling"
rule wasn't satisfied. Now I have to change the element order
to get it to compile. But, as a user, the change was completely
clear.
Yes, the order of the steps isn't the order of the flow... but I
don't care about that because I'm just experimenting with the
pipeline.
When the pipeline gets more complex, this is going to get worse. As
a user, I'm going to get angry with the WG for that constraint.
In all these messages, I'm trying to point out "good" reason why, when
the compiler *can* figure this out, we shouldn't have an seemly
arbitrary restriction for the user.
--Alex Milowski
Received on Monday, 2 October 2006 17:28:12 UTC