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

[Lou Burnard]

I wonder if the reason we're having this debate is because we're 
confusing two levels of interpretation which should be distinguished?

I am thinking about <del> and chums in the context *only* of manuscripts 
and other objects which don't have any XML markup natively in them. When 
we add the <del> we are identifying acts of deletion expressed in other 
ways.  So in Gaby's excellent example

<persName>Joannis Priskos<delSpan spanTo="#a"/>III</persName>, 
of<placeName>the
Tauric<anchor xml:id="#a"/>  Chersonese</placeName>

the problem is that the <persName> and <placeName> markup belong to a 
different layer of interpretation from the <delSpan>. If we had some 
mechanism for doing so, we'd want to be able to say that the 
</placeName> should magically be transformed into a </persName> after 
the application of the deletion. Suppose (as I suggested earlier) the 
manuscript had some defined set of conventions for distinguishing names 
of persons from those of places, say colour of ink. The conscientious 
scribe making the deletion would presumably in that situation want to 
indicate that the colour of ink used for Chersonese was now wrong -- and 
might even do so by retracing it in a different colour!

I dn't have a practical solution to offer I'm afraid, other than to 
enunmerate clearly the limitations and problems implicit in using 
@spanTo. Maybe Syd was right and it just is more trouble than it's worth.



On 22/08/11 12:22, Gabriel Bodard wrote:

> I'm not convinced that we can make this call in either direction, to be
> honest. The idea that a spanning element implied the deletion of all
> content and markup between the beginning and the end is problematic in
> any case where the deletion element is not a simple part of the
> hierarchy--that is to say in exactly those cases where<delSpan>  would
> be used. In terms of XML semantics, surely it's very problematic to
> delete the string "</persName>  abc<persName>" and assume the remaining
> opening and closing tags form a single element. (Indeed many graphical
> XML editors don't allow you to do this.)
>
> I also don't think that a case like:
>
> <persName>Joannis Priskos<delspan/>III</persName>, of<placeName>the
> Tauric<anchor/>  Chersonese</placeName>
>
> is as much of an outlier as is implied by James's admonition to the
> coder to avoid using delSpan to leave behind ill-formed XML. By
> definition it's going to happen: if it weren't, delSpan wouldn't be needed.
>
>> I think providing two deletions when there is a single line (without
>> indicating that these are fragmentary elements or something like that)
>> is encoding an untruth.
>
> This would only be the case if it is implied (which you're assuming it
> is, but I don't see why we have to) that a single<del>  element
> represents a single (and the entirety of) act of deletion/correction. If
> a single act of deletion effaces the last couple of words on each of
> four lines, we'd have to represent that with 4 del elements, I think,
> although it's a single act.
>
> Having said all this, I don't know that I have an answer to the question
> of how to interpret markup between a spanTo and its anchor. Maybe assume
> any completely contained elements are deleted, but any elements that
> close after the spanTo should be closed at that point, and any that open
> before the anchor should be opened at that point? Or expect the coder to
> document in the encodingDesc how any difficult cases should be handled?
>
>