Re: Remarques on W3C Editor's Draft 13 November 2007

/ Innovimax SARL <innovimax@gmail.com> was heard to say:
| Some more

Thanks again.

| Please do not preempt the future of XML Schema
| s/As per [W3C XML Schema: Part 2] or its successor(s),/As per [W3C XML
| Schema: Part 2],/

You don't want to say "or it's successors"? I suppose it's slightly
risky...

| === p:error ===
| I'm still concerned by the fact that p:error could not generate a
| c:error with all its attributes defined in 4.6.1.2 c:error
| May be we should whether remove some attributes from c:error or add
| some option to p:error of both

How does it not generate all the attributes?

<c:error
  name? = NCName       ==> name of the p:error step
  type? = QName        ==> p:error
  code? = QName        ==> code specified in p:error
  href? = anyURI       ==> URI of the pipeline that contains the p:error
  line? = integer      ==> line number of the p:error in the pipeline
  column? = integer    ==> column number of the p:error in the pipeline
  offset? = integer>   ==> offset of the p:error in the pipeline
    (string |          ==> description specified in p:error
     anyElement)*
</c:error>

| === p:hash ===
| what's the meaning of the option "value" ?

That's the value to be hashed.

| More than that, I strongly concerned by the fact that the hash value
| of an XML Infoset has never been defined (and by the way it would
| depend on serialisation option)

There's no infoset involved, p:hash hashes a string.

| In such a case, I think we should narrow the use to content of
| attribute or to string() value

All options are strings.

| === p:uuid ===
| Same concern as above

Which one?

| and in addition, not clear how it would (when the definition would be
| complete) from p:label-elements

I don't understand what you're asking.

| ==Ghosts==
| p:validate-xml-schema is still used in examples and even in Figure 1

Bah. Fixed.

| When you say "uncommon", which spec are you pointing ?

XDM, XPath 2.0, XML 1.0, XML 1.1, xml:id, XPointer, ...

That said, a little poking on the TR page suggests that they're still pretty
popular.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | ...it is significant that we are called
http://nwalsh.com/            | the 'information society' -- not the
                              | thinking society, not the deliberative
                              | society, not the society of reason and
                              | rationality.--Lloyd Morrisett