See Rick Jelliffe's SGML/XML Cookbook or David Megginson's Structuring XML
Documents ( http://www.xmlinfo.com/books/ )
>Anyway, given a lump of styled text with font changes, bolding etc, there
>seem to be several possible ways to map this.
>
>1) the obvious "HTML-style" of defining tags for <B> etc. and applying them
>inline, then adding the XSL definitions to match. This has a huge benefit
>of giving us HTML conversion in the same hit.
Just look at Word97's Save as HTML to see why this isn't trivial.
>2) some attempt to infer a document model, and defining the stylesheet on
>the basis of location (eg: the word "school" within paragraph 1). This
>feels very awkward but delivers cleaner text with separate styles.
Hard to do generically.
>3) conversion to RTF and encoding as same, on the basis that this will be
>parsable by other tools as an embedded format in future.
If you are happy to transport a presentational format around.
James