Errors like this:
  line 11 column 5 - Warning: <meta> element not empty or not closed

Looks like those come from an XMl parser.
Parsing HTML with an XML parser is like parsing FORTRAN with a CParser,
you get errors, but it doesn't mean that there is anything wrong with
the file.

  A piece of the generated html page:

That looks Ok except of course all teh xml namespace declarations 
xmlns:... are invald html attributes.


> With closing i also mean: <br/> (properly closed empty tag)

But as I said /> is XML syntax. It is not the syntax for an empty element
in HTML.

> Ok i understand, but html transitional and html strict are not xml
> languages? Or am i interchanging xhtml and html wrongly?

No version of HTML is an XML language. They are all defined via SGML
DTD. The XML versions are all called XHTML.

> So the conclusion of this is:
> IF your writing an html application that uses xsl transformation
> to output the html, then
> you have to change the xsl for each xsltransformer-implementation,

No that is not the conclusion at all.
You don't have to change your stylesheet but you may have to specify
that you want html serialisation in other (system specific) places.
The reason why XSLT doesn't mandate that xsl:ouput always has an effect
is that it allowes the result tree to be passed on as an in-memory treee
(or stream of sax events, or any other internal representation) This is
what happens in cocoon or if you run xslt inside of mozilla.
If the result tree is being passed on in such an in-memory format it is
never serialised to a linear document including tags so the hints on
xsl:output as to how to serialse the tree are never used.


> Is there no possibility to enforce the parser to use <br/> for
> empty tags?

You can force xslt to use <br/> syntax by using the xml output method,
but that is not correct html markup for a line break..

You have to decide what you want to do, generate html or generate xhtml,
and in the later case you need to decide if you want to make IE be able
to read the file as it has no built in xhtml support (unlike say mozilla
or opera which can render xhtml files).

The simplest, if you do not need xhtml in the document, is just to
generate html.

David