Re: Ignoring empty paragraphs

From: "L. David Baron" <>
Date: Tue, Apr 11, 2000, 9:45 AM

> On Tue, 11 Apr 2000 13:46:07 -0700, Matthew Brealey
> ( wrote:
>> > >it introduces a gap of about 1em, depending on your browser's
>> > >setting for margins on <p>.
>> >
>> > It does _not_ in Mozilla M14 "strict" mode.
>> > (yes, I did try exactly your line up here before posting :)
>> That's a bug - it doesn't ignore the Ps here:
>> so by any interpretation of the spec this is wrong.
> As I said in my original post [1], Mozilla's current approach is
> between the two I described.  I believe the current implementation is
> that the empty P elements are put into the DOM tree and the CSS layout
> code has special rules built in for HTML P elements.  (These rules
> don't lead to completely ignoring empty p elements, since a "<p></p>"
> in the middle of a block still does cause a line break).

Tasman similarly puts empty P elements into the DOM tree and has special CSS
layout code to "ignore" empty P elements.

Similarly, empty P elements take part in adjacent sibling selector matching in

> I'm not sure
> that one can argue that this is wrong, considering the vagueness of the
> statement mentioned in [1] from section 9.3.1 of HTML.


> I don't really like this behavior either.  However, without an
> authoritative statement from the authors of that statement about it's
> intent (as Braden mentioned in [2]), I don't think I'd be able to (or
> should) convince others to change this behavior.

I'll go beyond that - I'm fully convinced that the HTML4 spec intended that
empty P elements be ignored purely from a presentation perspective (face it -
there's tons of presentation related perspective in HTML4 - it's not just a
structural "markup" spec).

If it does turn out that the original intent was that empty P elements were
supposed to be ignored in the parser, then this is a bug in HTML4 spec which
needs to be fixed.  Fortunately, no such bug exists in XHTML.


Want standards? Have you validated your HTML today?

Received on Tuesday, 11 April 2000 13:33:23 UTC