Hi,
Lets say I have a MIXED tag
element
in my XML DTD with content that is going to be rendered as HMTL, as
well as a <bold> and <italics> tag.
I would strongly recommend that you don't call elements after
appearances unless you are designing a DTD for describing typography
(and even then probably not).
Be careful.
Specifically, does the whitespace in PCDATA of the MIXED tag get
rendered into the output html? Or is it a setting in the XSL
transforming document?
It's a pitfall. Yes, the white-space in mixed content does get preserved
in the HTML serialization in XSLT, but only if you use the default
setting of preserving space.
If you use <xsl:strip-space elements="*"/> to "tidy up" the output
serialization, irrelevant white-space is (correctly) suppressed between
elements in element content, and is still preserved in mixed content
*except* between adjacent elements (eg ...he said <italic>No</italic>
<bold>NO</bold>") where it is classed as a white-space-only node and
suppressed (omitted) so that you get...he said <i>No</i><b>NO</b>.
This is an unfortunate side-effect of XML being minimally processable
without a DTD or Schema (where element vs mixed content cannot be
foretold), and is one of the very few pieces of bad design in XSLT
because it happens *even* when the DTD is being used, which is precisely
the circumstance when it should *not* happen (when element vs mixed
content *is* foreknown).
You will be told -- by people who should know better -- that XSLT must
produce identical output whether run with or without a DTD/Schema, but
this is untrue (for example, default attribute values may cause
different output).
As web browsers ignore all excess white-space anyway (except in <pre>
elements), this shouldn't cause you any problems: just don't use
strip-space until you are more familiar with the rules governing
white-space nodes.
///Peter
