Re: Remarks

/ Innovimax SARL <> was heard to say:
| 1) In 5.8.1 Declaring atomic steps
| [[
| If p:log <> or
| p:serialization<>elements
| appear in the declaration of an atomic step, they will only be used
| if the atomic step is directly evaluated by the processor. They are ignored
| if the step appears in a pipeline.
| ]]
| 1.a) Please clarify the last sentence

Suppose we say that

  <p:delare-step type="ex:someStep">
    <p:serialization whatever parameters go here/>

Then if you run ex:someStep directly:

 $ xproc -run ex:someStep

the serialization parameters will be used. (Because that step *is* producing
the serialized results.)

But if you put <ex:someStep> in a larger pipeline, its serialization
parameters have no effect.

| 1.b) Please explain what happen in this context (declaration of atomic
| steps) when attributes amongst (@name, @psvi-required or @xpath-version)
| are used ?

The @name attribute is pointless, but not an error, on atomic steps.
I suppose @psvi-required and @xpath-version might be useful if the
step is being executed directly.

| For example, not sure to understand what should happen, if a pipeline in
| declared in p:library use  a  particular version, and the the declare-step
| in  a  pipeline that reference that  library use  one, and that the
| declare-step of this atomic step  use another one... ?

I think we have to say either that only the @xpath-version at the
"top" of the pipeline is considered (others are ignored) or that its
an error if the @xpath-version is specified in several places and is
not the same in all places.

| 3) Please make consistent 7.2.7 p:www-form-urldecode and  7.1.17
| p:parameters
| One is output a sequence of c:param document, the other is outputing a
| c:param-set
| Since order is implementation defined, please use c:param-set for both.

Ok. And if the output is a single c:param-set then it doesn't have to
be a sequence. I like that.

| Furthermore, I would add a constraint saying that no more than one c:param
| with the same QName shoud be outputted in this c:param-set (it could
| recalled in The c:param-set element)

I think that was the intent of the sentence "The step resolves
duplicate parameters..." but I added a clarification.

| 4) In 7.1.17
| [[
| Note
| Since the parameters port is *not* primary, any explicit
| p:with-param<>settings
| must include a port attribute with value
| p:with-param<>,
| per the last paragraph of Section 5.7.4,
| "p:with-param"<>
| .
| ]]
| Please replace "with value p:with-param" to "with value parameters"

I simply removed the "with value" phrase. I think it was redundant.

                                        Be seeing you,

Norman Walsh <> | Happiness is a how, not a what; a            | talent, not an object.--Herman Hesse

Received on Thursday, 24 April 2008 13:38:30 UTC