W3C home > Mailing lists > Public > public-html@w3.org > April 2007

Re: Proposing <indent> vs. <blockquote>

From: David Dailey <david.dailey@sru.edu>
Date: Fri, 20 Apr 2007 13:56:40 -0400
Message-Id: <>
To: public-html@w3.org

At Fri, 20 Apr 2007 17:12:55 +0200 Alexander Graf wrote:

[Lots of nice stuff, that I generally agree with ... ]

 >They also have used nested  <blockquote>s to increase indentation. 
How would >that work using attributes? I  think it would confuse 
users even more.

I agree, as I've mentioned, lots of teachers who use HTML think of it 
as ALL presentational. HTML is to many just a way to present 
information. To me it is a way of partitioning a rectangle. Would 
that those partitions were not all rectangular, and that the 
dimension of the thing being partitioned were higher than two! 
Perhaps an n-D generalization of Voronoi diagrams would make for a 
good browser topology. What fun! The screen rendering is then just an 
exercise in projective geometry -- with some weird boundary 
conditions owing to "semantic" concerns.

But your question above, about multiple indentations and ease of 
slapping something together using blockquote for an upcoming class 
(FrontPage indents by inserting <blockquote>) made me ponder another 
situation: The use of tables as presentational vehicles.

I don't know if the argument I am about to make holds water or not -- 
it may stem out of my simple ignorance about CSS. But I suspect there 
may be things for which <table> as a presentational device 
outdistances the performance afforded by styling.

<table cellspacing="2" cellpadding="2" border="1">
     <td rowspan="2">A</td>
     <td colspan="2">B</td>
     <td rowspan="3">D</td>
     <td colspan="2">F</td>
     <td rowspan="2" colspan="2">G</td>
     <td rowspan="2">H</td>
     <td colspan="2">J</td>

Above, we have drawn a rectangle on the screen and partioned the 
rectangle into 10 smaller rectangles. (I have an ongoing offer of 
extra credit to anyone who can enumerate the number of distinct 
partitions thusly representable by a <table> with exactly n <td>'s 
using only colspan and rowspan attributes. The reader is encouraged 
to submit such solutions, but off-list please.) The table wizards in 
Macromedia Homesite, MS FrontPage, and Adobe Photoshop (there, it is 
called an image slicer), each provide ways of enabling the author to 
carve rectangles into sub-rectangles. Apparently the use case is 
strong enough that those three companies (now down to two) pursued a 
solution (and here I thought Adobe was fairly forward looking). I use 
the stuff for a) partitioning images into user-defined regions for 
subsequent morphing between two images and b) rapid prototyping of 
the onscreen aspects of interface design (the classic 2D stuff).

Is there some easy way to build such an entity using styles which 
would render <table> obsolete for this purpose? If so, I suppose it 
is time for me to do some more reading on CSS.

BTW what I don't have in mind is this: 
in which a page is filled with a domino tessellation of image 
thumbnails of differing aspect ratios. It uses no tables, but the 
scripting here is just a bit intense for your average user of an HTML 
editor. (I guess it's now too late to patent that, huh? oh well.)

David Dailey
Received on Friday, 20 April 2007 17:57:27 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:19 UTC