Re: Nesting a flat XML structure

On 29/10/18 21:52, ian.proudfoot@xxxxxxxxxxx wrote:
> Going a little off topic, but the concept is relatively simple. Many
> writers don't make the best use of their word processors. Maybe lists
> are manually indented with bullets inserted from a character palette.
> Titles may be 'Normal' text with character overrides for font size and
> weight. 

Indeed they are. This was the starting-point in my session at the XML
Summerschool on how to deal with Word documents as a source for XML in
publishing.

> Careful analysis of many documents showed that there are between
> eight and ten properties that have the most effect on the output for
> character styles and paragraph styles. This is presented as an
> override code in a format that is very compact but also possible for
> anyone to understand. The combination of any correctly defined style 
> name plus its override code gives us a key that can be used for
> mapping to elements in the output.

Very ingenious b have you published this? I'd be interested to compare
it to the dozen or so areas I investigated when looking at the use of
editing software for structured documents
(https://cora.ucc.ie/handle/10468/1690)

> This works well when there is some inherent logic to the implied 
> structure of the source document. Less so when no regard has been
> given to sensible style use.

That's the key, of course. The problem of getting authors to adhere to
stylesheets is a lost cause of many years (in most cases: there are a
few exceptions). In effect, as Wendell and Tommie put it, "the author
sees it as *his own job* to invent the schema"B9

///Peter
-- 
B9 Piez, W., & Usdin, T. (2007). 'Separating Mapping from Coding in
Transformation Tasks'. XML Conference. Boston, MA: IdeAlliance.