Re: questionable html constructs

Dan Connolly (connolly@pixel.convex.com)
Fri, 15 Jan 93 18:21:10 CST

>>o What is the significance, if any, of <p>'s within a <pre> section?
>
>They are allowed, but I don't know that they should be interpreted,
>since the new-lines are also significant. I'd say that they should be
>interpreted as ADDITIONAL line feeds; the people who set their code
>up otherwise can just fix their html. Ideally, you'd want to realize
>that there's a <P> with every new-line, and if so, ignore one or the
>other. That's not really a legal treatment of SGML, but it should
>produce nice results. Of course it encourages people not to fix
>their HTML.

This question is out of the jourisdiction of the SGML standard:
it's a question of how we're going to interpret things and display
them.

But, unlike the normal typeset paragraphs where ignoring a newline
here or there is no big deal, the writers of PRE sections want
some assurance as to what the output will look like.

It's on the todo list. I suggested we ignore <p> within <pre>,
but that looks like a bad idea. So now the current thinking,
I believe, is to render it as a line break. This makes those
man pages that end in <p> _and_ newline double space, but
we'll put the onus on the provider to fix that.

>>o What is the significance, if any, of more than one <p> in a row,
>> particularly on the same line? And, how about the same situation in
>> a <pre> section?
>
>It seems to be legal HTML, so I guess you just throw in a totally empty
>paragraph.

I agree (again: this is not an SGML question, but a WWW question).

>>o Is '&lt' ('&lt;' without the semicolon) a valid construct?
>
>No.

Now this is an SGML question, and you're wrong: it's legal.

>>o Can raw <'s and >'s be in a <pre> section?
>
>No.

Again, this is an SGML question, and you're wrong.
<PRE>
Is a < b?
</PRE>
is perfectly valid markup.

Dan