W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > August 2007

Re: New draft: 10 Aug

From: Alex Milowski <alex@milowski.org>
Date: Mon, 13 Aug 2007 12:37:31 -0700
Message-ID: <28d56ece0708131237v520c58f2ne3cff6173a765b78@mail.gmail.com>
To: public-xml-processing-model-wg@w3.org

(fixes are checked in but not online yet)

On 8/10/07, Innovimax SARL <innovimax@gmail.com> wrote:
> Remove unneeded sequence="no" in p:add-attribute


> Please add sequence="yes" to output port=result in p:add-xml-base or
> remove it from input port="source"

I think this should not allow sequences on input to be consistent with out
policy of not creating steps that operate upon a document but do implicit
iteration.  I've removed sequence="yes" from the source.

> Rephrase second bullet
> [[
> If the element is the document element and *there is no explicit has
> no xml:base attribute*, an xml:base attribute is added with the value
> set to the element's base URI.
> ]]
> see between stars


> It seems option=all need to have a default value to 'no', and
> option=relative also

I think it would be better to have the 'relative' option default to
'yes'.  I most cases
the combination of all='no' and relative='yes' should result in a document with
a single xml:base attribute on the document element with an absolute URI of
the document's location.

Otherwise, this is fixed.

> --
> In p:directory-list
> Why is filter a Regex ?

It needs to be a pattern of some sort so you can match "*.xml" or
"*.xhtml", right?
The F&O document from XPath 2.0 already specifies how patterns as regular
expressions work.  Otherwise, we'd have to find another specification to point
at and I don't see a reason to do that since we already use the F&O document

> --
> In p:insert
> Now that there is no default for position, It needs to be required="yes"

I agree.  Fixed.

> --
> In p:label
> Please remove the first line (redundant) and the unneeded "and the step fails"
> [[
> If an existing xml:id value conflicts with a previously generated
> value, the step fails.
> It is a dynamic error (err:XC0006) if an existing xml:id value
> conflicts with a previously generated value and the step fails.
> ]]

Ah... forgot to delete the old text.  Fixed.

> --
> In p:namespace-rename
> Please rephrase this sentence
> [[
> If the from option is the empty string, or is not specified, then
> elements and attributes in no namespace are renamed. If the to option
> is the empty string, or is not specified, then elements and attributes
> in the specified from namespace are renamed into no namespace.
> ]]

I've made the language a bit cleaner.

I found a problem in this step that I fixed in that if you rename the
no namespace you'd probably rename every attribute and that's
probably not what you wanted.  I've fixed this by adding a
'elements-only' option.

Also, there was a dynamic error for this step that wasn't called out
properly that I fixed.

> --
> In p:parameters
> All the c:parameter points to p:parameter, please fix it
> as with p:document/p:documentation, I suggest to find another name for this step

Not fixed yet... need to check with Norm about where he wants to describe those
elements as they appear many places.

> --
> In p:rename
> I think we should narrow the error XC009 to target only the validity
> of the name option
> [[
> It is a dynamic error (err:XC0009) if the specified name is not a
> valid name or if the renaming would introduce a syntactic error into
> the document (i.e., if it would create two attributes with the same
> name on the same element).
> ]]
> to
> [[
> It is a dynamic error (err:XC0009) if the specified name is not a valid QName
> ]]

The restriction of having two attributes of the same name needs to remain.  I've
made it a separate error.

I've added the QName bit to the error.

> For the matter of validity, if we want to keep such level of detail,
> we need to provide the same for each step (which could be a huge task)
> We already know that if the document is not namespace aware well
> formed, it will be a XD0001 error (note en passant that XD0001 is not
> as precise at it should)

Our documents are based on infosets and so are namespace aware.

> --
> In p:set-attributes
> you need to provide a default value for match or make it required="yes"
> Note en passant, that all match option are not consistant in each steps
> --

I made it required as all the others are required.

> In p:serialization
> "The encoding must support the values UTF-8 or UTF-16."
> Please replace "or" by "and" (UTF-8 should be mandatory)

I think we should just require UTF-8 and skip all the others.  Any
expected to work in many ideographic languages is going to support
UTF-16 or UTF-32 or some non-standard encoding.  UTF-8 is sufficient to
encode those languages even though it results in a much a larger document.

> "The omit-xml-declaration must support be supported. If the value is
> not specified or has the value no, an XML declaration must be
> produced."
> please correct "must support be supported". Also, Is the
> omit-xml-declaration="yes" mandatory (your phrasing is not enough
> clear)

Correct?  That's the language directly from the serialization spec.

--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language

Bertrand Russell in a footnote of Principles of Mathematics
Received on Monday, 13 August 2007 19:48:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:32:44 UTC