[osis-core] unbalanced <q> elements

[Chris Little]

Todd,

>>I'm marking up the TEV and could use some advice.
>>
>>We've got <q/> for marking balanced quotations.  And we've got
>><milestone type="cQuote"/> for marking continuation quotes.
>>
>>1) If I'm marking quotation levels, how should I indicate this on
>>continuation quote milestones?  Should they be linked with their <q/>
>>parent elements (perhaps with a new cID attribute, similar to the sID
>>& eID attributes)?
> 
> 
> I can appreciate the academic case where a continuation quote mark
> occurs within a nested quote, but is this a case you have actually run
> into or are you just trying to be explicitly clear in what you are
> encoding?

The TEV uses single and double quotes.  Double quotes are sometimes 
embedded within double quotes to mark single words.  Regardless of how 
deeply embedded they actually are, I use the single/double quote 
distinction to decide the value of the level attribute.

That said, there are instances where both single & double quotes cross 
paragraph boundaries and require continuation quotes.

> I would rather see <milestone type="cQuote.1"/> and <milestone
> type="cQuote.2"/> than add an attribute to <milestone> that is tailored
> specifically to continuation quote marks.

I think that would work for me.

> We could also specify that one of the following:
> 1) <milestone type="cQuote"/> ONLY applies to the outer most quote OR
> 2) <milestone type="cQuote"/> applies to the level="1" quote OR
> 3) <milestone type="cQuote"/> applies to the <q> container that it is
> the most immediate child of.
> Would any or all of these solve the problem you are facing?

These would all be unable to handle multiple levels of continuation 
quotes, like the TEV has.


>>2) If I have unbalanced quotes that are not continuation quotes, how
>>should I mark them?  Perhaps with a <q/> that has an sID or eID but
>>without a matching <q/> with the same value?  Specifically, I have
>>some close quotation marks inside of notes that lack matching open
>>quotation marks.
> 
> This exact issue is one that I ran into before and we did not get
> resolved in the OSIS meeting last summer.
> 
> We have said that it is invalid for an element with and sID to not have
> a matching eID and visa versa.  
> 
> Your suggestion seems to be the most natural.  Possibly we could allow
> <q sID=""/> where the empty value indicates that there will be not
> matching eID and vica versa.
> 
> Alternatively we could add <milestone type="unbalancedLeftQuote"/> and
> <milestone type="unbalancedRightQuote"/> (with better attribute names of
> course).
> 
> We should try to come up with a solution the upcoming meeting.  

Any of these solutions would work for me.

--Chris