Re: Load Component

/ Alessandro Vernet <avernet@orbeon.com> was heard to say:
| On 2/1/07, Norman Walsh <Norman.Walsh@sun.com> wrote:
|> Under the status quo, I think it's a violation of the spec. I think
|> that's true even if you use an extension attribute. So I think this
|> has to be a different component unless we provide some sort of hook
|> for it.
|
| I wouldn't go that far. Consider:
|
| <p:step name="load">
|    <p:param name="href" value="http://www.google.com/"/>
|    <p:param name="x:tidy" value="true"/>
| </p:step>
|
| This makes it clear that x:tidy is an extension parameter. I don't
| think we should prevent implementations from innovating by adding
| features on top of the pipeline language as long it is clear that
| those are extensions which might not be supported by other
| implementations.

Now I *really* want two kinds of parameters:

  <p:load x:tidy="true">
     <p:param name="href" value="http://www.google.com/"/>
  </p:load>

Consider

 <p:step name="xslt">
    <p:input port="stylesheet" href="foo.xsl"/>
    <p:param name="href" value="http://www.google.com/"/>
    <p:param name="x:tidy" value="true"/>
 </p:step>

Does the presence of x:tidy on this component signal some magic
behavior? Does it mean that if foo.xsl produces HTML it'll be tidied
into XHTML? How can any processor know that it's encountering an
extension attribute. What if I don't want any extension behavior but I
want the x:tidy parameter of the foo.xsl stylesheet to have the value
"true"?

At least with

 <p:xslt x:tidy="true">
    <p:input port="stylesheet" href="foo.xsl"/>
    <p:param name="href" value="http://www.google.com/"/>
 </p:step>

My processor can know that the author was attempting to indicate some
extended behavior.


                                        Be seeing you,
                                          norm

-- 
Norman Walsh
XML Standards Architect
Sun Microsystems, Inc.