That would complicate certainly the CERN browsers, in that if they see a <p>
they always do the same thing. They don't have mechanisms for being
context-sensitive about these things -- they just do the right thing with
good HTML and don't bomb out with bad html.
I'd prefer us to say that the effect of <P> within <PRE> is undefined, or that
it is not recommened but if used is a newline. I'd like Thomas to take them
out of his man pages, but butil then it isn't a disaster, its just double
spacing. I feel that it would be a kludge to say they had no effect within
<PRE> ... it would look weird to historians.
--
I am prepared to change the <XMP> behaviour to be standard SGML for the sake of
it even if the example secitions in our documents end up moving up or down a
line as a result.
Tim