Re: Closing ACTION-2099

Chris Lilley:
>  ACTION-2099
> Review the corrections to the mistakes in the attribute index and
> report back
> I reviewed the appendix - particularly wrt the schema - and did not
> see any issues that needed correcting for publication.
> It would be *nice* but not required, to change the values on
> properties so that 'inherit' came in a consistent position in the list
> (eg, last) but that is an editorial matter.


> It would be good for the attribute names in the table to be linked
> to the corresponding section of the document, like properties are. I
> understand the attribute appendix is auto generated; where is the raw
> data stored? i wouldn't mind updating it so the attribute appendix has
> links to the definitions. Again, this is an editorial nicety and is
> not required for publication.

There isn’t a file mapping attribute names to spec location.  There’s
master/elements.txt and master/properties.txt for element and property
definition links.  Someone could make one for the attributes, but this’d
need to map (element, attribute) pairs to spec location (possibly with a
wildcard for the element, to handle common attributes).

Chris Lilley:
> Of course, as soon as I sent that, i saw that id is listed twice, as
> ID and as NCName. The two entries should be combined. It should say,
> for the type
> <ID> | <NCName>
> and the attribute name should link to the definition, at
> struct.html#idAttrs

I wonder if the way id/xml:id is defined in the schema is related to
ISSUE-2009?  Anyway, I’ve special cased id in the table generation
script so it comes out as one line.

Cameron McCormack ≝

Received on Tuesday, 9 September 2008 12:37:05 UTC