[
Lists Home |
Date Index |
Thread Index
]
> Derek Denny-Brown wrote:
>
> Even _with_ 20/20 hindsight I don't think anyone's devised
> a way to unambiguously and reliably identify mixed content
> without a DTD or some other external device that provides
> vocabulary-specific information.
It allway depends if you (1) _can_ have a DTD or if you (2) _can_ avoid
mixed content.
For (1) this is not always possible and if I remember correctly the concept
of wellformedness was one of the strong points for XML during ist
developent, especially as it makes writing parsers easier. It did even have
an effect on SGML.
For (2) if you have data centric xml you normally _can_ avoid mixed content,
however in document centric xml this is not possible.
> For all the quirks and seemingly-bizarre whitespace rules
> in SGML, it's one of the few markup technologies where
> whitespace handling tends to Just Work (as long as you follow
> best practices like "avoid pernicious mixed content").
True SGML is flexible, extensive and strict, but this also makes it
expensive ...
Axel
|
