Re: Sub-pipelines

Hi,

[Many apologies if my comments are out-of-whack due to my prolonged 
absence: I've read all the minutes, but they never capture everything.]

Alessandro Vernet wrote:
> In fact, we could consider that we have a pre-defined component called
> "pipeline" that just runs pipeline, just like "xslt" runs a
> stylesheet:
> 
>     <p:step name="pipeline">
>         <p:input name="pipeline" href="expandandvalidate.xpl"/>
>         <p:input name="schemas" href="a.xsd"/>
>         <p:input name="document" href="mydoc.xml"/>
>         <p:output label="ok1"/>
>     </p:step>

I agree with Norm's observation that always having to define reusable 
pipelines in separate documents makes things harder than they should be. 
I have two other arguments for why we shouldn't have a pipeline 
component as the only method of reusing pipelines:

(1) It makes it harder for GUIs, other helper applications, and users to 
identify the reusable parts of a pipeline.

(2) It opens the door to dynamically generated pipeline specifications, 
and I don't think we want to go there (do we?).

I don't think that we can or should stop pipeline engines from making a 
pipeline-interpreting component available, but I think we should have a 
native method of reusing subpipelines.

In fact, once a pipeline is given a name then it should be usable just 
like any other component. So once you've defined the 'expandandvalidate' 
pipeline, you should be able to invoke it using the standard syntax for 
invoking components, such as:

   <p:step name="expandandvalidate">
     <p:input name="schemas" href="a.xsd" />
     <p:input name="document" href="mydoc.xml" />
     <p:output label="ok1" />
   </p:step>

Cheers,

Jeni
-- 
Jeni Tennison
http://www.jenitennison.com

Received on Monday, 20 March 2006 12:49:58 UTC