On Mon, 2 Feb 2004, Daniel O'Donnell wrote:
> Hi Roberto,
> I'm hoping this is the recommended way, personally. It is certainly how
> I do it. All these pairs that seem to rely on position cause problems,
> of course. You can always use the corresp and ID attributes to make the
> relationship between expansion and contraction explicit if there is an
> opportunity for confusion (as there might be for example if you had a
> string of distinct contractions and abbreviations in a long word).
Wouldn't it be better to wrap the <abbr> and <expan> in a <seg> or
<ab> or something? So instead of:
or indeed if encoding at such a level:
I guess what I'm concerned about is the opportunity for confusion
that you note with a string of disctinct contractions/abbreviations
in a long word. Using <ab> (if that is indeed the correct thing
to use?) you at least know that everytime you come across
an ab[@type='abbrev'] that you have a children, and a potential
choice. Whereas with mann<abbr>«</abbr><expan>um</expan> you
might get this confusion you suggest if the immediately following
<expan> belongs to a different abbreviation. I wouldn't really
be worried about it as long as the encoder is consistent, of course.
This is, of course, the perfect situation for Lou's <choice> (or
whatever) element hopefully appearing in P5.
Dr James Cummings, Oxford Text Archive, University of Oxford
James.Cummings at ota.ahds.ac.uk +44(0)1865-283296