Re: Leave only p:xpath-context

>* p:iteration-source allows sequence and optionnaly a @select
It only allows @select because there's no "select" option to this step.

Providing such an option will not only eliminate this as argument, but will
also simplify XProc, as <p:for-each/> will become as intuitive as
<xsl:for-each/>. That is, you'll be doing:
<p:for-each select="//chapter">

>* p:xpath-context allow zero or one document and no @select
Why not simply disallow the omission of a binding? <p:empty/> is still there
to specify an empty context node if that is what the pipeline author wants.

>* p:viewport-source allow one and only one single document and no @select
This is pretty much where I suggest that p:xpath-context goes. The only
thing that will have to be clarified in that case is that a sequence of
documents is allowed only when p:xpath-context is inside p:for-each.

-----Original Message-----
From: mozer [mailto:xmlizer@gmail.com]
Sent: Saturday, August 04, 2007 5:55 PM
To: Vasil Rangelov
Cc: public-xml-processing-model-comments@w3.org
Subject: Re: Leave only p:xpath-context

I do not agree

There is sensible differences between each
* p:iteration-source allows sequence and optionnaly a @select
* p:viewport-source allow one and only one single document and no @select
* p:xpath-contect allow zero or one document and no @select

It seems that those elements are sufficiently different to not be merged

Xmlizer

On 8/4/07, Vasil Rangelov <boen.robot@gmail.com> wrote:
>
> For simplification to pipeline authors, I think the p:iteration-source and
> p:viewport-source should be removed in favor of p:xpath-context.
>
> There seems to be no reason for not doing so. They all share the same
> function (to create a binding for a compound step), the same content (a
> single binding, pointed either with p:document, p:pipe, p:inline or
p:empty)
> and everything.
>
> The specification does describe the semantics of p:iteration-source a bit
> differently though, so I guess that the it will either have to provide a
> generic definition (which practically means removal of the "select"
> attribute) or describe the semantics of p:xpath-context in that special
> case.
>
> Regards,
> Vasil Rangelov

Received on Saturday, 4 August 2007 16:52:13 UTC