[tei-council] span span span span span span span span glorious span

[Lou Burnard]

As I said on the call, I am currently  working through tcw18 (the 
genetic editing doc) and constructing from it a proposed set of 
additions to P5. In due course you can expect to see a long list of 
requests for comment and minor clarification here, as well as the 
document itself.

However. I'd like to raise here and now one specific point which I have 
just hit, since it's not one we've discussed explicitly, and which has 
fairly major implications.

This is the correct usage for the attributes @target and @spanTo. A 
number of the new elements proposed in the working paper are meant to 
operate in "stand off" mode, and use these attributes to indicate the 
bits of transcript to which they apply. (examples include metaMark, 
undo, redo). The document is a bit vague about how these attributes are 
supposed to operate in combination, if at all, and the examples cited 
are mutually contradictory.

I'd like to propose the following set of rules.

1. @target behaves as elsewhere in the Guidelines -- it's a URL, which 
can point to one or more whole elements, or xpath-defined substrings

2. @spanTo is also (as elsewhere) a URL., but it must point to a 
*single* element.

3. it is erroneous to supply both attributes on the same element.

4. if only @target is supplied, the passage/s affected are all the 
element content fragments indicated by the URIs supplied, treated as a 
single unit

5. if only @spanTo is supplied, the passage affected is the sequence of 
content fragments that begins immediately following the element 
concerned and finishes  immediately preceding whatever element is 
indicated by the @spanTo attribute.

6. if neither attribute is supplied, the markup is erroneous

It's important to get this, if not right, at least consistent, as we 
have a number of other places where @spanTo has been invoked as The 
Solution.

I'd also appreciate help with the following conundrum

Suppose we find

<p>blah <delSpan spanTo="#delEnd"/> blah </p>
<p>blah <anchor xml:id="delEnd"/> blah </p>

Clearly two of my "blah" content nodes are being deleted. What about the 
tags? In other words, if I write a processor to act upon the markup and 
produce a new XML representation of it, should it produce

(a) <p>blah blah </p>

or

(b) <p>blah</p><p>blah</p>

I am pretty sure I know the right answer, but I am not sure I know how 
to express the algorithm such a processor should apply.