Michel Rodriguez <mrodrigu@[...].org>  writes:

>  On Tue, 25 Jul 2000, Tylman Ule wrote:
>  
>  > However, I came across the same problem pointed out by Jay J.
>  > Simplifying his code, I think the problem is, in a nutshell, that empty
>  > elements are treated as if they have a closing tag, and, as a result,
>  > their content appears to their right. The code below results in the
>  > following XML output:
>  
>  > my $doc = <<'_EOT_';
>  > <?xml version='1.0' standalone="yes"?>
>  > <document>
>  > <some_tag/>
>  > <some_tag></some_tag>
>  > </document>
>  > _EOT_
>  
>  This is illegal I think. I don't believe a tag can be both empty AND not
>  empty in the same document. I am surprised that Expat did not croak on it.

 3.1 Start-Tags, End-Tags, and Empty-Element Tags

  [...]

  If an element is empty, it must be represented either by a start-tag
  immediately followed by an end-tag or by an empty-element tag. An
  empty-element tag takes a special form:

  [...]

  Empty-element tags may be used for any element which has no content,
  whether or not it is declared using the keyword EMPTY. *For
  interoperability*, the empty-element tag must be used, and can only
  be used, for elements which are declared EMPTY.

[emphasis mine]

This reads to me that if you are validating (not Expat), and the
parser comes across a start-tag and end-tag pair for an element that
is declared EMPTY or an empty-element for an element that is not
declared EMPTY, then it should be reported as a warning.

A non-validating parser treats a empty start-tag/end-tag pair and an
empty-tag as effectively the same thing.  Expat, for example, calls
both the start-element and end-element callbacks.

In your message to xml-dev, you mention about an element having
content in one place and an empty-tag being used for the same element
in another place.  A validating parser would catch that as above but a
non-validating parser won't care.

  -- Ken