W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > December 2002

Syntax: positive review

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Thu, 19 Dec 2002 12:07:38 +0100
To: w3c-rdfcore-wg@w3.org

One substantive change proposed.

8 critical changes
17 near-critical

 - see comment on parseTypeOther 7.2.20 - my recollection is that the WG was
unsupportive of this production but left it in, my suggested correction
reflects this - I am not sure where we got to really.

 - Dave uses the word "interpreted" a few times. I think its fine but you
may differ.

Syntax review


OK for Last Call, subject to one substantive critical change,
and further critical changes.


I was much more positive about section 2 this time.
I am not sure if it was the changes that have been made
since I last reviewed it, or whether I was simply in
a better mood. It was fairly clear and correct.
I believe it adds to the document.

Substantive change

Add the following text to ?section 5.1?
 Within RDF/XML documents it is not permitted to use XML
 namespaces whose namespace URI is either:
 + the RDF Namespace URI reference concatenated with additional characters.
 + the XML Namespace URI reference concatenated with additional characters.

(The current text permits:

  <DF xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#R">
    <andom rdfa:bout="eg:a"
     xmlns:rdfa="http://www.w3.org/1999/02/22-rdf-syntax-ns#a" />

as the RDF graph of one triple:

<eg:a> rdf:type rdf:Random .

I think the suggested text is a better backwardly incompatible
clarification of M&S than the current text.

Further comments

Each is marked
  *C* critical
      MUST be considered/changed for last call.
      Typically these are minor incorrect statements.
  *N* near-critical
      SHOULD be considered/changed for last call.
      (I realise the series editor has a preference for
      minimising change at this point).
      Typically these are misleading or incomplete statements.
  *S* spelling/type
      will need correcting sometime, this are sufficiently
      trivial that I see no reason not to do them now.
  *E* editorial
      suggestions for improvement
      I have put less work into these.

I proceed in document order.


*E*  Do we want the recommendation to continue to refer to
     the issues list?
*E*  Repetition of rdf:nodeID and rdf:parseType="Collection"
     could be avoided.

*E* first para, last sentence
      Does not read easily, could be deleted.
*E* second para
      repeats SOTD, charter and WG not relevant to rec,
      suggest delete.

*C* first para:
    [[ Nodes are RDF URI References, RDF Literals or are Blank Nodes and for
encoding in syntaxes may be given a document-local, non-RDF URI References
identifier called a Blank Node Identifier
    Nodes not Blank Nodes is the subject of the relative clause.
    Suggest split into two sentences:
    [[ Nodes are RDF URI References, RDF Literals or are Blank Nodes.
      Blank Nodes may be given a document-local, non-RDF URI References
identifier called a Blank Node Identifier

*C* second para
    XML NS maps Qnames to URI refs not RDF URI refs
    suggested text: (2 deletions of "RDF" and losing hyperlinks).
RDF/XML uses XML QNames to represent RDF URI References. The namespace
prefix part of all QNames is associated with a URI Reference as defined in
XML Namespaces [XML-NS]. The RDF URI Reference represented by a QName is
determined by appending the local name part of the QName to the URI
Reference associated with the namespace prefix part of the QName

2.2 Example 3
*C* Add:
    (this example fails to show that the blank node is
     shared between the two paths)
*N* In above, add [[see 2.10]]

2.3 Example 4
(related to *C* iin Ex 3)

*N* (this example does show that there is a single blank node)

*C* First para: [[requires]]=>[[permits]]

*E* Second para (3 changes insert [[correct]]
    [[3]]=>[[4]] [[the complete]]=> [[a complete]]
    [[    This could be done for any of the correct complete graph
examples from Example 4 onwards but taking the smallest Example 6 and adding
the final components, gives a complete RDF/XML representation]]

*E* First para last sentence misleading. Suggest
    [[The xml:lang="" form indicates the absence of a language identifier.]]

*E* Suggest add reference to Example 3.
*E* Suggest base Example 11 on Example 3.

*N* At end of first para add:
    [[Property attributes and rdf:nodeID attribute
      are not permitted on such property elements.]]

*N* Delete word [[Blank]] in section heading
*N* Delete word [[blank]] in first sentence
*N* Add sentence at end of first para
    [[If the node is a RDF URI reference this is shown
    using an rdf:resource attribute (see 2.4).
    If the node is blank then an optional rdf:nodeID
    (see section 2.?? ) may be used.

2.13 Example 14
*S*? marvelous (twice) I think its "marvellous"
     maybe not google votes 89600 to 42000 for your

*N* Second para
    [[    This provides an additional check since the same name can only
appear once in the scope of an xml:base value (or document, if none is
given), so
    I could misread as permitting
  <a xml:base="eg:a" rdf:ID="b" />
  <a xml:base="eg:a" rdf:ID="b" />

   but not allowing
 <t  xml:base="eg:a" >
  <a rdf:ID="b" />
  <a rdf:ID="b" />

  I prefered your later text on the same topic. How about
  Each (rdf:ID attribute value, base URI) pair has to be unique in an
RDF/XML document and so this provides an additional check]]

*E* First para
    suggest [[about]]=>[[connecting]]

*E* Second sentence was difficult.
   The identifier for the triple is a RDF URI reference
   formed from by: concatenating # with the rdf:ID attribute
   value, considered as a relative URI reference and
   resolved against the in-scope base URI.
  perhaps that's not much better.

*N* Before [[See constraint-id]] suggest replace
    [[as the rdf:bagID attribute]]
    [[as the rdf:bagID attribute and of values
     of the rdf:ID attribute used in section 2.14]]
   (perhaps needs more wordsmithing).

*S* straight after Ex 20 "for a rdf:bagID" => "for an rdf:bagID"

*E* "that lists the statements" => "that lists the reifications of the

  See first issue listed

*S* First sentence "arcs" => "arc"

*C* 2nd para [[resolved to]] => [[resolved with]]

*S* 3rd para [[may may]]=>[[may]]

Considering para "Information items ... XML Literals.."
*N* delete [[and]]
    => [[...XML Literals do not follow ...]]
    rephrase by moving this para up to be part of list Document Information
Item , Element Info Item, Attr Info Item, Char Info Item. This would need
more work and is probably inappropriate for this round now.


*C* under attributes

The xml:lang stuff as written does not conform with XML, bother.
The correction above probably fixes it.

The problem is that you allow binding a different prefix to the XML
namespace or worse something like:

<a xlan:g xmlns:xlan="http://www.w3.org/XML/1998/namespacelan">

In XML (which does not know about namespaces) xml:lang is an 8 character
string, not a QName. I think your text is fine, but needs my first
correction above.

A much more substantial correction, which corresponds to the behaviour of
ARP, is to keep the QName as a <URIref, localname> pair until later in the
processing than the model in the document. This could have been done but is
not appropriate now, particularly given that the innocuous additional
paragraph suggested fixes the problems.


*N* Last two paragraphs
    Should not be NOTEs since they have normative content.

*N* Last Call should not refer to next version of this WD

*N* Suggest
    Replace last paragraph by:

    Within a single element of an RDF/XML document,
    it is not permitted to use  an unqualified attribute
    and a qualified attribute with the same
    local name and namespace name being the
    RDF Namespace URI reference.

6.1.6, 6.1.7
*N* Suggest adding a note like that in 6.1.8
  [[Note that the value of the identifier may need processing to meet the
syntactic constraints of N-triples]]

6.1.8, 6.18
*C* There are two sections 6.1.8

*N* Suggest delete note about combining chars. This is stated in Concepts,
and the issue is suffficiently minor that that IMO suffices.

*N* also in the other section 6.1.8 same comment

*E*  In table for anyURI suggest delete [[legal]] superfluous.

*C*  Last phrase:
    [[but no state is expected to be preserved]]
    it is not true since some things are scoped by the document

*S* [[are are]] ==> [[are]]


*S* [[URIS]]=>[[URIs]]


*S*  [[If the element content c is not an empty, then]]
     delete [[an]]
*E*  Delete all of quoted phrase - it works in the empty case too.

*N* Add para at end of section
New documents SHOULD NOT use this production.
Applications MAY choose to warn when this production is seen in a document.

*N* Reorder the three first levelbullets under the text:
    "The following are done in any order:"
    + For all propertyAttr attributes ...
    + Add the following statement ..
    + If event n ...

    This clears any possible misreading of "for each statement above" since
the only statements in the same section above are the ones intended.
    It has no other impact since the bullets are explicitly unordered.

*E* Delete last phrase [[ that now allows ... serialized]]
    (It is slightly misleading, if quote out of context)

Not reviewed.
Received on Thursday, 19 December 2002 06:08:41 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:54:56 EDT