Re: Term="yes|no"

Hi Yves,

Yves Savourel wrote:
> Hi Felix, 
> 
> - Actually you are pointing out something I didn't touch on before because it seemed too obvious: I would add the value "no" to the
> local term attribute as well.

But you did not list the necessary place in the spec to be changed - or
did I miss s.t.?

> If we have "yes|no" available in global rules they should be available locally as well. So one could
> do somethin like this:
> 
> <its:termRule selector="//kw" term="yes"/>
> ...
> <p>This is some <kw its:term="no">text</kw></p>
> 
> 
> - I'm not sure if I understand your point when some data categories do not have override.
> They seem override to have to me:
> 
> for ruby, while it makes *much less* sense to use override because the nature of the information is not a flag but a specific text,
> technically you can do it too:
> 
> <its:rubyRule rubyText="Click this image to see a larger version" selector="//@alt"/>
> <its:rubyRule rubyText="World Wide Web Consortium" selector="//image[@src='w3c.png']/@alt"/>
> 
> And the same goes for localization information.
> No?

I'm more worried about overriding and its interplay with inheritance,
see below.


> 
> In any case, I would think any "flag"-type data category should have a way to override.
> 
> I guess to see the issue from a different viewpoint: how do we justify that term cannot have a "no" value (locally and globally)?

To me, the issue of inheritance is difference for terminology: if we say
<termdef its:term="yes">...<em>..</em>...</termdef>
the content of <em> does not inherit the its information its:term="yes"
from <termdef>, but it is rather a part of the ITS term. This is
different also for ruby:
<its:rubyRule rubyText="Click this image to see a larger version"
selector="//rubyBase"/>
does mean that <rubyBase> is a ruby base element. This property *does
not* inherit to the content of ruby base. In the case of localization
information, translatability and directionality, the case is different:
here we have really CSS like "inheritance".

> Currently to cancel a termRule from in an external file in a document instance you have to comment it out, which is not a good
> option since such external file may be used by different document where the same rule is needed.
> If I recall correctly the only reason we removed term="yes|no" and made limited local term to "yes" was because we thought the cases
> for "no" simply did not exist. Which is clearly not the case.

I'm still worried this is a substantive change. Of course it depends on
how we see it ...

Cheers,

Felix

> 
> Cheers,
> -yves
> 
> 
> 
> 
> 
> -----Original Message-----
> From: Felix Sasaki [mailto:fsasaki@w3.org] 
> Sent: Wednesday, August 30, 2006 12:52 AM
> To: Yves Savourel
> Cc: public-i18n-its@w3.org
> Subject: Re: Term="yes|no"
> 
> Hi Yves,
> 
> Sorry for being late in this discussion. I have some concerns in this
> change:
> 
> - It disconnects the global usage of the terminology data category with the local one. In the latter, we have only term="yes".
> - It introduces a new functionality for global rules of overriding a "this is a term" rule, which again is not available locally.
> - I think the comparison to xml:lang regarding overrides, which Martin introduced, is not appropriate, since xml:lang is only used
> locally.
> - You wrote "One should be able to override a previous rule that says a given element is a term.", but I'm not sure if this is
> absolutely necessary. Translatability, directionality and elements within text use overrides, but the other data categories don't.
> 
> I'm also concerned that this change, esp. the disconnection between global and local, is rather substantive and not appropriate
> during last call.
> 
> Again, sorry for being late and my concerns.
> 
> Felix
> 
> Yves Savourel wrote:
>> Hi all,
>>
>> I have an action item to list the changes needed in the WD to add term="yes|no"
>> http://www.w3.org/2006/08/23-i18nits-minutes.html#action01
>>
>> Here they are:
>>
>> A) Add term="yes" in <its:termRule> in the examples 12, 13, 15, 17, 19 and 27.
>>
>>
>> B) In section 6.4.2: "is realized with a termRule element with a mandatory selector attribute."
>>
>> Would become:
>>
>> "is realized with a termRule element with a mandatory selector attribute and a mandatory term attribute with a value 'yes' or
> 'no'."
>> (or whatever more consistent formulation matches the one Christian has 
>> come up with during the last edit of the data cat
>> definitions)
>>
>>
>> C) In section 6.4.3: The addition of term="yes|no" in the termRule's attributes list ODD definition.
>>
>>
>> I think that is all.
>> -yves
>>
>>
>>
>>
>>
> 
> 
> 
>