Re: Re: Use Case 5.27: Integrate Computation Components (MathML)

On 7/12/06, Alex Milowski <alex@milowski.org> wrote:
> I'm using 'name' consistently where you are using 'label' and
> 'from' where you are using 'name'.

Alex,

OK, I think I understand your point. Correct me if I am
misinterpreting what you just said:

1) You value the consistency we get by naming attributes on
<p:input>/<p:output> based on the direction of the data flow. Hence
your solution has 2 attribute names, for instance:

<p:input name="stylesheet" from="my-stylesheet"/>

Which means: "Take the data from 'my-stylesheet' and send it to the
input named 'stylesheet'."

<p:output name="result-of-my-transformation" from="result"/>

Which means: "Take the data from the output 'result' and name it
'result-of-my-transformation' so I can reference it with this name in
this pipeline."

2) I value the consistency we get by using the same attribute name for
the name assigned to inputs/outputs by the component. Hence my
solution has 3 attribute names, for instance:

<p:input name="stylesheet" from="my-stylesheet"/>

Which means: "The component defines an input named 'stylesheet'. For
this step, take data from 'my-stylesheet' and feed it to that input."

<p:output name="result" label="result-of-my-transformation"/>

Which means: "The component defines a output named 'result'. Label
this output 'result-of-my-transformation' so I can use that label to
reference that output in this pipeline."


Both approaches seem valid to me. Would there be another option which
is better than those exposed here? If not, and this is an open
question to everyone here, which options seems the most consistent or
intuitive?

Alex
-- 
Blog (XML, Web apps, Open Source):
http://www.orbeon.com/blog/

Received on Thursday, 13 July 2006 17:36:03 UTC