Hi George.
On 21/10/05, George Cristian Bina <[EMAIL PROTECTED]> wrote:
> 1. There is an inconsistency (spelling error) between the definition and
> use of QName, it is defined as
> {Qname}
> --^----
> and used as
> {QName}
> --^----
Ouch. OK.
I have the XML in subversion now. I'll fix it this weekend.
>
> 2. The pattern mark should be followed by ": ", now pattern is
> immediately followed without any separator by the element name.
OK
>
> 3. Some descriptions start with an upper case letter, others with a
> lower case letter. They should be changed to be consistent.
...... Grrr. OK :-)
>
> 4. The description of the attribute pattern says:
>
> Pattern matching an attribute
> Pattern matching attributes
+1.
>
> 5. The description of the element pattern says:
> matches a named element
> Again, a name-class defines a set of names so the description should be
> changed to read:
>
> Pattern matching elements
+1
>
> 6. The except pattern is marked with a name-class marker instead of pattern.
My bad. Done.
>
> 7. The grammar pattern has the namespace declaration
> (xmlns='http://relaxng.org/ns/structure/1.0') added to its description,
> is that necessary?
I think so.
It's rare in Eric's book, so a reminder won't hurt.
>
> 8. Again the grammar pattern specifies the ns and datatypeLibrary
> attributes, but there is nothing special with them in the grammar
> pattern, that are present in all Relax NG elements.
Mmm.
This is where I'd define them for local ns's and
documentation, comments etc.
That's why I added them here.
Not 'wrong' .... I'm tempted to leave them (aide memoire?)
> One approach will be to define a model like common-attributes and add
> that to all the elements, but that will increase the complexity. Another
> approach is to specify (as the document already does) that these two
> attributes are present on all elements.
Optionally present? Clearer - (Not currently present George?)
I'll try and add that.
>
> 9. A spelling error in the name description, there is "withsingle"
> instead of "with a single"
+1
(You remind me of a braille transcriber! Finds all the tiny errors
that are near invisible to me :-)
>
> 10. Shall special symbols like {QName}, {text} and {empty} link to their
> definition? I think that this will be good if it is possible.
? To the notes at the top? ..... Could do, but I use a document
like this in print form, where the links wouldn't be much use?
I'll do it, and create an HTML form too.
>
> 11. The notAllowed patter contains in its description also in what
> elements it can be present. This is inconsistent with the other
> descriptions, so is there any reason for that?
It's 'odd', different, negative to all the others being positive?
>
> 12. The nsName name-class does not require a ns attribute.
I think it does.
>It means any
> name in the namespace specified by the inherited ns attribute so the ns
> attribute does not need to be specified here. It is present in the
> document as a general attribute. Removing it causes also model and the
> description to be changed, maybe for description we can have:
> Any name in a namespace.
> or
> Any name in the namespace specified by
> the inherited @ns attribute.
Can we have another opinion on this please?
I think the ns attribute is required.
>
> 13. The description of the oneOrMore pattern has a spelling error:
> The children of this elements may be used one or more times
> ----------------------------^
+1
>
> 14. I do not really understand the description of parentRef:
> Extends the scope of a named pattern to that of a pattern in
> the parent grammar
> I proposed:
> Defines a reference to a named pattern in the parent grammar
1. I've never used it.
2. I'm taking the text from Eric's book.
Resolution from other readers please?
>
> 15. The combine attribute is present in 2 description, for the start
> pattern and for the define pattern. In one place it is:
>
> May specify combine as choice or interleave.
>
> and in the other
>
> @combine='choice'| 'interleave'
>
> I think it will be good to have a similar presentation in both places.
Yes.....
Abbreviated or 'full English'?
I think I prefer the abbreviated form.
>
> 16. The define element should be marked as a pattern.
+1
>
> 17. The value pattern does not have a datatype so I think the datatype
> word should be removed from its description:
> Matches a text node against a datatyped value
> ------------------------------^^^^^^^^^
OK... syntax (or me being lazy with the xml!)
>
> 18. The definition of the content models is not consistent with how the
> names are specified, the name-class is specified as "name-class" and the
> pattern as "&pattern;".
Relic of former format.
I'll delete the entity inference.
> Also any reason to mark the name-class choice in the name-class content
> model but leave choice not marked as pattern in the pattern content model?
I'll add (as pattern) in the pattern content model.
Thanks George.
Appreciated.
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk
------------------------ Yahoo! Groups Sponsor --------------------~-->
Fair play? Video games influencing politics. Click and talk back!
http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/2U_rlB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/rng-users/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
