We are working with the CES-DTD in our project "Deutsches Referenzkorpus"
at the Institut für Deutsche Sprache in Mannheim.
We want to encode different kinds of headings, especially in newspaper texts.
Now we have a problem with cross titles. Each time we want to tag a cross
title, we have to begin a new division:
<div n="0" type="article">
<head type="main">Jugendliche unter sich</head>
<div n="1" type="section">
<div n="2" type="section">
But this approach has some difficulties. Without the definitions of the CES
DTD we would not encode a new division; because text1 and text2 and text3
are at the same level.
If there are many crosstitles, specially when there are sequences of many
text snippets in newspapers, we get a very complex division structure that
seeems not correct.
Has anybody some suggestions how we could solve this?
Institut für Deutsche Sprache
mail: [log in to unmask]