Why is XML hierarchical? 
That is something I wonder every time I look into XML dumps of certain software. The last two candidates were
OmniOutliner 3 Professional and
Apples Pages.
While OmniOutliner, a generally excellent software, is a tool that works with hierarchical data structures and even stores a very nice OPML file the XML used for export plugins seems to throw away the idea of embedded tags for styles and other information about the data. Why is the style of a piece of text defined by a tag just in front of it?
Even more chaotic is the situation with Pages. While at the user level a very nice and very easy to use text processing tool it stores its files in an XML file that is cryptic and ignores lots of features you get by using XML. The first thing that struck me was that all styles are stored right with the text itself. If you remove the styles Pages chockes heavily and refuses to open the document. It would have been too easy to separate the styles from the actual document. If so we could change the look in a matter of seconds, but now its more like a Word document.
Second, all text information is stored in paragraphs and the structure of the document is determined by paragraph styles. So a headline is not declared as such, but simply a paragraph that has the attributes that format it like a headline. While I can understand the benefits of such an approach, the drawbacks are that it is much more difficult to work with such a data format.
And why am I ranting about this? I want to use OmniOutliner as a tool to write my documents. Its a clean start with headlines and subtitles. I can fill in text as notes and always focus on the topic. I want the benefits of Pages too. However, for now I will have to settle with converting OmniOutliners format to an intermediate XML I used previously to write documents and convert them to LaTeX. This works great, a while business plan had been written this way once.
