W3C home > Mailing lists > Public > www-style@w3.org > September 2005

RE: CSS vs tables for layout

From: Mark Moore <mark.moore@notlimited.com>
Date: Wed, 21 Sep 2005 15:48:41 -0700
To: <www-style@w3.org>
Cc: "'David Woolley'" <david@djwhome.demon.co.uk>
Message-ID: <E1EIDRK-0006Nx-HI@lisa.w3.org>

Wow, I am surprised by the argument against collecting objective data.

Joseph, let me counter David's opinion with one on the other side:  I
believe the effort to quantify the costs of table based layout vs CSS driven
layout would be very beneficial.

Throwing some objective, fact-based light on the subject can only be
goodness.  The fact there is such a gap between what should be and what is
demonstrates the research is relevant.

I suspect you'll find there are quantifiable differences in the development
costs and especially the support cost of a website once it's initially
deployed.  Abstracting the display from the content yields major savings as
designs grow and change under real world market demands.

To answer your follow on question, I can't think of a relevant alternative
that competes with the CSS vs table layout design problem.  (If they're not
using one of those two techniques, they're doing something else.)

With that said, my suspicion is the table vs CSS layout maybe strongly
influenced by the choice of development tools.  You may want to factor that
in as you research the space.


> -----Original Message-----
> From: www-style-request@w3.org [mailto:www-style-request@w3.org] On Behalf
> Of David Woolley
> Sent: Wednesday, September 21, 2005 2:45 PM
> To: www-style@w3.org
> Subject: Re: CSS vs tables for layout
> > I had not considered addressing "CSS table display types".  I'm assuming
> > that browser support for this is minimal as practitioners rarely (if
> > ever) address it in discussion.
> That's not a safe assumption to make.  Whilst there may be some truth in
> it the main reasons are:
> -  the vast majority of practitioners who use tables for layout have never
>    read the W3C specifications or the browser vendors' technical
> documentation.
> -  the majority of them have not bought into the idea that web pages are
>    structured documents and only want to use HTML/CSS as a means of
> achieving
>    a particular visual goal (in some cases the practioner may have bought
> the
>    idea, but it is unlikely that their client has);
> -  for such practioners, HTML table mediated layout achieves what they
>    want and there are plenty of examples to copy, so, in the absence of
>    a desire for structure, there is no incentive to do things properly.
> My impression is that most web page coding is done by copying examples
> from other web pages, a little from cookbooks, which tend to be
> compilations of questionable practice discovered by looking at how
> web pages are coded, and only very few actually understand the languages
> that they are using.  (This list is extremely unusual in that everyone
> here will have had some exposure to the actual specifications.)
> > My intent is to consider the costs and benefits of different layout
> > development methods for practioners.  In addition to considering the
> That sounds like working around current browsers.  That's definitely
> off topic here.
> > "CSS table model" as one of those methods, are there any other
> I believe that is the only method of the two that you mentioned that is
> within the scope of this list.
> > methods/problems/considerations you feel are relevant, especially as
> I think some consideration should be given as to why people have got
> into the situation where they need to transmit several different documents
> every time they send an HTML resource to the user.  I think navigation and
> logos should be linked from the head section and merged into the display
> by the browser.
> > they relate to gathering data on the topic.
> Re gathering data, you mustn't make inferences from what hasn't been done
> into what is impossible.
> > Do you feel such an analysis, or variation on this is relevant to
> > practitioners?
> I doubt it.  Most are totally unaware and those that are aware tend to
> already be aware of the problems.
Received on Wednesday, 21 September 2005 22:52:00 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:20 UTC