Paragraphs Elements as Containers

Robert Seymour robert_seymour at il.us.swissbank.com
Sun May 7 14:52:23 PDT 1995


In parsing a paragraph element, omniweb does not distinguish between the use  
of a paragraph tag pair as a container and a paragraph tag stadalone as a new  
paragraph descriptor.  The two uses are (for example):

	<p>This is my paragraph in a container.</p>

	This is my paragraph using an ending tag only.<p>

In each case shown above, the paragraph break should be inserted after the  
end of the sentence.  However, OmniWeb will insert the paragraph break  
wherever the <p> tag occurs, and ignore the presence of the </p> paragraph  
container ending tag.  As a consequent of this parsing, OmniWeb may induce  
paragraph breaks in inappropriate situations.  A simple example would be:

	<ul>
		<li><p>A list item.</p>
	</ul>

(Now, the studious among you will note that this HTML instance does not  
conform to the SGML DTD for HTML 2.0 (or 3.0), but I just wanted an easy  
example for us to look at.  I'm sure you can imagine other cases with similar  
results.)  In this case, OmniWeb would draw a bullet element, break the end  
of line and place the text "A list item." on the next line of the screen.   
What it has done is to interpret the <p> as the end of a paragraph, ignoring  
the </p>.  This is contrary to the directions for using paragraph elements in  
Raggett's and TBL's RFCs/DDs (as I understand them).  Are there any plans to  
fix this and have OmniWeb do a look-ahead for a </p> before inserting a  
paragraph break at a <p> element?

On an extension to this question, will OmniWeb parse align arguments to the  
paragraph tag (e.g.: <p align="justify">)?  While not as useful as text  
wrapping around images, it should be fairly trivial to add on and does not  
require hacking the text object.

Regards, Robert


--
Robert Seymour			 seymour at il.us.swissbank.com
Swiss Bank Corporation		 (NeXTMail, MIME, PGP accepted)

Internal Web: <URL:http://sbcweb/~seymour/>
Internet Web: <URL:http://www.reed.edu/~rseymour/>


More information about the OmniWeb-l mailing list