Re: Directionality Scope/inheritence issue (same as translatability)

Hi Yves,

Yves Savourel wrote:
> Hi Felix,
> 
> 
>> In addition to the problems you mentioned in your mails 
>> about directionality and translatability, there is another 
>> one: the global rules don't take the default selections in 
>> the instance into account.
> 
> I'm not sure if I understand the paragraph above. What are the "default selections in the instance"?
> 
> I'm guessing it would (for translatability for example): "element content is translatable, attribute values are not". But the global
> rules take that into account... (so I probably don't understand).

at http://www.w3.org/TR/its/#selection-defaults-etc , we define
"default" selections. I think for translatability this means if the have
<p its:translate="..">...</p>
the value of the translate attribute is related to "Textual content of
element, including content of child elements, but excluding attributes",
no matter what the value is.

for directionality it means if we have
<p its:dir="rtl">..</p>
it means "Textual content of element, including attributes and child
elements"

Now, a global rule like
<its:dirRule select="//p" its:dir="rtl"/>
selects only the <p> elements, not their attributes. The same is true
for a global rule like
<its:translateRule select="//p" its:translate="..."/>
this rule does not say "attributes are not translatable": it just does
not talk about attributes.


the problem with the global rules which select s.t like
"//p[trans='yes']" is that you have to add all kinds of "implicit"
interpretations of XPath, to achieve the interpretation at
http://www.w3.org/TR/its/#selection-defaults-etc . The other problem is
inheritance, which you mentioned before.

The third problem is that people probably want to be able to use e.g.
@trans attributes like local its:translate attributes: to describe
translatability for some specific cases. The global cases should be
described with normal global rules, e.g. as you did for XHTML. However,
if we use global rules for interpreting @trans, @trans will have the
same precedence position as other global rules. I guess user want it in
the precedence position of its:translate (which would be possible with a
pre-processing step which transforms @trans to its:translate).

However, my impression was if we have in the XML document s.t. like
<p trans="..">...</p>
we want to be able to say "trans has the same meaning like *local*
translatability", that is, like
<p its:translate="...">...</p>

I am not sure yet about the syntax for s.t. like that. Maybe
<its:translateRule its:translate="yes" realized-as="//@trans['yes']"/>

or for directionality
<its:translateRule its:dir="rtl" realized-as="//@dir['rtl']"/>

Does this make sense?

- Felix

> 
> Can you elaborate?
> 
> Thanks
> -yves
> 
>