[osis-core] Revisiting unbalanced quotes

[Chris Little]

Patrick Durusau wrote:
> Guys,
> 
> Working back through my email and noticed Chris's post on unbalance quotes.
> 
> 2 Cases as I understand it:
> 
> 1. Nested continuation quotes (can someone give me a reference to an 
> example in the TEV) where the issue is now to link a continuation quote 
> to its "parent".

The decalogue, Deut 5.7.

> 2. A quote in a note that has the beginning quote mark but not a closing 
> quote. (not at all sure that would be meaningful to a reader but assume 
> someone has an example of this)
> 
> Question on #2: If the beginning but not the end of the quote is marked, 
> how do you know where the quote ends? The reason I ask is that with some 
> attribute value as a trigger, you could mark the beginning and end of 
> the quote and set your stylesheet to simply not output a closing quote 
> mark for that type of quote. Yes, it is in a quote container but that 
> does not mean that it has to have quote marks at both ends.

The end of the quotation would occur at the closure of the container 
that contains the starting quotation mark.  In the case of a <note>, 
probably a </catchWord> tag.

Importantly, we would need two types for these quote containers that 
only render one of their marks: one for an unbalanced left quote and 
another for an unbalanced right quote.

> On #1, just curious if marking the quotes using milestones, and using 
> the osisID, with the "!" extension, would be enough to link them 
> together. Not real sure how one can tell what is a continuation quote of 
> what in the circumstances described but then I have not seen the 
> example. Same mechanism as I suggest for #2 in terms of getting a quote 
> mark only at one end or at both.

I don't think the end of the quotation will be as easily determined for 
case #1 as for #2.  However, using the same system as suggested for #2 
would not be semantically correct for #1.  Take the example below:

<p>He said, <q level="1" sID="l1"/>blah blah, she said <q level="2" 
sID="l2"/>blah blah.</p>
<p><q level="1" type="continuing"><q level="2" type="continuing">blah 
blah</q></q></p>
<p><q level="1" type="continuing"><q level="2" type="continuing">blah 
blah</q></q></p>
<p><q level="1" type="continuing"><q level="2" type="continuing">blah 
blah</q></q><q eID="l1"/><q eID="l2"/></p>

By introducing actual <q> containers for continuing quotation marks, we 
end up suggesting that there are additional levels of quotation.  We're 
sacrificing accurate semantics for presentation.  I believe in Dallas, 
we were moving towards a solution that made use of cQuote milestones 
with ID/IDREF attributes to link back to the start quote tags, such as 
below:

<p>He said, <q level="1" sID="l1" ID="l1"/>blah blah, she said <q 
level="2" sID="l2" ID="l2"/>blah blah.</p>
<p><milestone type="cQuote" IDREF="l1"/><milestone type="cQuote" 
IDREF="l2"/>blah blah</p>
<p><milestone type="cQuote" IDREF="l1"/><milestone type="cQuote" 
IDREF="l2"/>blah blah</p>
<p><milestone type="cQuote" IDREF="l1"/><milestone type="cQuote" 
IDREF="l2"/>blah blah<q eID="l1"/><q eID="l2"/></p>

--Chris