Re: more p:output thoughts

/ James Fuller <james.fuller.2007@gmail.com> was heard to say:
| On 6/8/07, Norman Walsh <ndw@nwalsh.com> wrote:
|> / James Fuller <james.fuller.2007@gmail.com> was heard to say:
|> | 5.5 p:output Element
|> |
|> | p:output can contain a p:document which in turn is defined as 'A
|> | p:document reads an XML document from a URI.'....is this an oversight
|> | of some sort? Is there any related error condition for this type of
|> | scenario?
|>
|> The construction
|>
|>   <p:output port="foo">
|>     <p:document href="someURI"/>
|>   </p:output>
|>
|> causes the content of someURI to appear on the output port "foo". I
|> expect it to be very, very uncommon. Used mostly, if ever, in p:catch
|> steps to produce constant error output.
|
| yes, but from a users point I would argue that expected behavior from
| such a construction would be 'save the output to anyURI'....willing to
| bet a pint that this will become a very common FAQ indeed.

Indeed. Actually, I think that our very recent decision to forbid
completely empty compound steps removes all the practical value in
allowing p:document inside p:output so I think I'm going to argue that
we remove it.

It used to be the case that you could imagine saying:

  ...
  <p:catch>
    <p:output port="result">
      <p:document href="someURI"/>
    </p:output>
  </p:catch>

But that's not legal anymore so I don't think there's any value in
having p:document or p:inline in p:output.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Old and young, we are all on our last
http://nwalsh.com/            | cruise.--Robert Louis Stevenson