Re: SPACE...

> From: Michael Sperberg-McQueen <U35395@UICVM.UIC.EDU>

> I had been thinking that all such separators except NBSP should also
> be subject to XML's white-space normalization rules (reduction of any
> span of white-space characters to a single SPACE -- i.e. U+0020 --
> unless white-space preservation is turned on)

> We were not planning to level NBSP, since one common use
> for it is to try to prevent such white-space normalization in
> specific cases

My quick re-reading of the WD doesn't answer the following for me:

Is white space normalization done "before" or "after" entity
expansion.  In other words, if a text entity starts, ends, or consists
of white space characters, does white space normalization occur (1) for
consecutive whitespace within the entity replacement text and/or (2)
for the concatenation of the whitespace preceding the entity ref and
the whitespace at the beginning of the replacement text?  Likewise, if
a character reference is to a whitespace character (other than NBSP or
any other for which ws normalization is defined not to occur anyway),
does ws normalization occur wrt all the replacement texts?