<PARA>
This text is <EMPHASIS>more</EMPHASIS> readable than
<UNDERLINE><EMPHASIS>either <SUPERSCRIPT>example</SUPERSCRIPT></EMPHASIS>,
and is easier for a human to process and follow.</UNDERLINE></PARA>
which is not markedly easier to read than this
<PARA>
This text is <EMPHASIS>NO</> less readable than the previous example and easier
to read than
<UNDERLINE>either of the <EMPHASIS>running-text<SUPERSCRIPT>examples</></>,
should it be necssary for a human to process and follow it.</></>
and, in either case, significantly better than running text with strictly
width-constrained line-breaks. if you really have to read marked up source,
you're working against the odds unless it's the shape of the text body, not
the content of the end tag, which matters.
the more general point is that, the argument for long tags <em>for the purpose
of human readability</> has little credibility, since one should place minimal
standards on the tools which present the marked up source to do so in a way
(ie with meaningful justification) which, as a side-effect, renders the end
tag content redundant.
i don't have to do with many editors, but the one which i do use (visual page)
tends to enforce such ("pretty printing") conventions. which i think was the
correct decision and well worth the extra implementation effort.