Dave Pawson wrote:
> Comments inline.
>
> On 26/07/07, Oliver-Rainer Wittmann - Software Engineer - Sun Microsystems
>
>> Thus, I propose to introduce the following new element and new
>> attributes:
>> - new element <office:annotation-end> with the attribute
>> office:annotation-end-id with type <ref name="ID">
>> - new optional attribute office:ending-at with type <ref name="IDREF">
>> for existing element <office:annotation>
>
> -1 to the offic:annotation-end
>
> I'd prefer a 'matching pair'
> e.g. annotation-start and annotation-end as the element names.
> or use an attribute to declare the end?
I didn't choose 'matching pair' of start and end, because of the
following reason:
An implementation, which supports ODF 1.1 will not lose the annotation,
which annotates a part of the document. Such an application will still
read the <office:annotation> element - it only ignores the unknown
attribute office:annotation-end-id and the unknown element
<office:annotation-end>.
Backward-Compatibility was the reason not the introduce new elements
<office:annotation-start> and <office:annotation-end>.
>
> I'm guessing you're using ID and IDREF as the type?
Yes.
> You don't say what they point to? I.e. where the match is expected to
> be for testing.
Value of attribute office:ending-at points to value of
office:annotation-end-id.
>
>
> More general comment.
> With insertion and deletion, the extent of the 'overlapping markup'
> starts to become
> significant?
That's correct.
>
> I guess it would help if they all followed a consistant pattern?
> There are prior works
> http://www.idealliance.org/papers/extreme/proceedings/html/2002/Tennison02/EML2002Tennison02.html
>
>
>
>
> Its becoming popular too. Eg. rng extension for validation of
> overlapping markup?
> http://2007.xtech.org/public/schedule/paper/46
>
>
> regards
>
