[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: XML Schemas: Best Practices
- From: Rick Jelliffe <email@example.com>
- To: firstname.lastname@example.org
- Date: Sat, 06 Jan 2001 16:36:52 +0800
I'm not really sure if this applies, but I am very wary of the idea of
abstract "containers" unless they are treated as degerenate skeletons (i.e.
a skeleton with no head and an implied body grouping its contents): the
abstract structure runs throughout XML, from the bottom (attributes=head,
content=body) to the top (most DTDs seem to end up with a metadata section
if they evolve long enough.
So when thinking about adding information items to abstract containers, I
think it is best practise to first figure out if the intended information
item belongs (either for semantic/coherence reasons or operation reasons to
allow stream-processing or) in a head section that the container may not
have previously had.
Following this line of thought, I guess that it would be best practise for
public, extensible containers (not being presentation-oriented) to always
have some provision for a head element (the body element does not always
need to be explicit). This might be an optional abstract element at the
start of the content model, for example.
One advantage of this is that it can allow localized metadata, during
cut-and-paste of a branch, it provides a parking space for information that
otherwise would be lost from the original context (given that we are still
along way from handling fragment-exchange context specifications during