W3C home > Mailing lists > Public > www-style@w3.org > June 2007

Re: Stylings only possible with Tables

From: Spartanicus <mk98762@gmail.com>
Date: Tue, 26 Jun 2007 09:46:15 +0100
Message-ID: <n2m-g.ltf1831eoqbrjntmv2rtau3tkof9tl3oso@4ax.com>
To: www-style@w3.org

Daniel Beardsmore <public@telcontar.net> wrote:

>> Regardless of whether or not a new CSS method allows implementation
>> algorithms to be reused, any new CSS mechanism would at best take a year
>> to specify ...
>
>Which is worse than the rest of CSS 3 how?

It isn't. To be honest I was rather optimistic with that 1 year at best
estimate, from conception to REC is likely to take considerably longer.
Things move slowly in W3C land.

>> about five years to be implemented ...
>
>Good grief. Either CSS is already in a far worse shape than I realised, or 
>you're severely over-estimating the complexity of basic layout. If it takes 
>*FIVE* years for people to write some simple layout code, then CSS needs 
>taking out, shooting and starting over.

Consider for example the CSS3 selectors spec, it reached CR status in
2001, it has only recently seen the first full implementation and it is
nowhere near to being supported by a large number of clients used on the
web today.

>Five years ... I don't know if browsers have crap coders, or CSS is so 
>unwieldy that it will take that long to figure out how anyone is supposed to 
>build on it.

The incentives to implement aren't there. The clients today that make
some effort get very little from it other than a warm fuzzy feeling and
applause from a small group of CSS advocates. It is of little use to
real world web authors, they have to code for the lowest common
denominator. CSS3 selectors serve a very minimal purpose today for use
in user stylesheets. Not so for implementing CSS3 Advanced layout.

> > and it will take at least
>> another five years before the use of legacy browsers has diminished
>> enough before such a mechanism can realistically be used by authors.
>
>Again, this is different to existing CSS 3 proposals in what way?

Again, it isn't.

>Go take a 
>look at the existing Advanced Layout module. It's not been altered in a year 
>and a half, it's full of "Are we even going in the right direction with 
>this?" questions, and you're telling me we can't slot in something simple, 
>straightforward, pragmatic and logical to get basic page layout away from 
>tables?

The reasons why many casual authors find tables easy to use for layout
are directly related to why they are in fact bad for layout. Any new
solution wont have both the "ease" of tables without the drawbacks, it
will be one or the other. Use tables if you want the ease, we don't need
something new.

It is regrettable that CSS2 didn't offer a basic method to do layout,
but that's water under the bridge now, we can't change the past. The
practical consequences of this shortfall are limited provided that
authors don't waste their time trying to use CSS2 for something that it
doesn't really cater for: columnar layouts. Sadly many authors do just
that: waste time trying to get CSS2 to do what it wasn't designed for.

Non columnar layouts are perfectly possible with CSS2, albeit that this
takes a fair bit of skill, partly due to an overly complicated spec,
partly because of browser bugs (IE mainly), but mostly due to the fact
that creating a good layout that doesn't exhibit any of the possible
pitfalls simply isn't easy.

Quality layouts are difficult, both to spec and to author. Casual
authors aren't going to grasp that, they should use tables as tables
despite the inherent drawbacks are likely to result in a better quality
layout when compared to an attempted "CSS layout" made by the same
casual author.

What we do need is a method to do layout for more sophisticated authors.
That method will be complicated, it is a complicated issue to address
comprehensively.

>I really hope you do NOT speak on behalf of sane Web developers, as tables 
>are being mutilated and violated for the sake of layout. Accessibility ends 
>up down the drain and it leads to painful server-side template coding.
>
>WE MUST GET OUT of the tables rut. NOW. No use using "well, tables already 
>work" as an excuse. Bullshit. When I want to tell people to stop using 
>tables, what do I tell them to use instead? Floats?
>
>This is my real sore point. I want to evangelise decent Web development but 
>it really pains me to realise that even I can't get straightforward designs 
>to work in CSS 2, let alone trying to move people away from tables.

Regrettably many people who advocate moving away from table layouts are
part of the problem because they have a flawed understanding of the
issues. The solutions that they use regularly cause more problems than
they solve.

-- 
Spartanicus
Received on Tuesday, 26 June 2007 08:45:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:51 GMT