W3C home > Mailing lists > Public > www-html@w3.org > December 1998

Re: Proposition...

From: Nir Dagan <nir@nirdagan.com>
Date: Tue, 29 Dec 1998 00:38:06
Message-Id: <>
To: "Inanis Brooke" <alatus@earthlink.net>, "w3c html" <www-html@w3.org>
At 14:57 28/12/98 -0800, Inanis Brooke wrote:
>If all of us are HTML developers, as well as avid web users, we know that
>many of the mega-sites whose layout is dependent upon tables take forever to
>load on a modem. While the latest web browsers render individual table cells
>upon recieving the </td> tag, there can still be a lot of HTML text in a
>single cell, giving the appearence that the user's connection to the HTTP
>server is dead.

An already existing solution from the content provider side is 
to use CSS positioning for layout rather than TABLEs. This usually 
does not require loading the whole content for calculating column 
widths. Also users on slow connections can turn off stylesheets for 
faster rendering.
A solution from the client could be to render TABLEs serially. Opera 
three and a half has this option. It simply renders every TD or TH as a P
when tables are turned off.

>To fix this, I have a proposition.
>I have wished for quite a long time that I could have a table with cells
>whose only data is a link to text which comes later in the document. I've
>included an example.txt file which shows what I'm talking about.
>This could dynamically speed up page rendering for modem users, 

I can't see how it does so. Clearly, browsers who do no recognize 
the import attribute and the export element will render an empty
table and a serial display of the content of the export element.
On the other hand a browser aware of these features will have to wait 
loading the last export element to calculate the table's dimentions.


Nir Dagan
Assistant Professor of Economics
Universidad Pompeu Fabra
RamonTrias Fargas 25
08005 Barcelona


"There is nothing quite so practical as a good theory." 
-- A. Einstein
Received on Monday, 28 December 1998 18:36:11 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:49 UTC