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

Re: We are where we are.... but how did we get here?

From: Orion Adrian <orion.adrian@gmail.com>
Date: Thu, 30 Jun 2005 00:33:43 -0400
Message-ID: <abd6c80105062921332d0c529f@mail.gmail.com>
To: www-style@w3.org

> > For example I would love someone to write an explanation on the W3C website
> > behind the box model the W3C choose to adopted?
> 
> I imagine this is an attempt to produce a simple but complete model.
> The difference between product marketing specifications and standards
> is that the former don't have to specify the awkward cases, whereas
> the latter have to do so, even if to say that the result is explicitly
> undefined (although, as authors don't read specifications, having
> explicitly undefined behaviour can be unsafe on the web).  The result
> is that standards aim to minimise the worst case complexity, whereas
> commercial specs only worry about the easy cases (e.g. the ones the
> salesman demonstrates).

But the trick I see is to cover as many cases as feasible and optimize
for the common cases. I feel that CSS covers most cases, but doesn't
optimize well.
 
> > Why is a second pass unworkable? (computing power has increased
> > significantly since 1998... I can play Doom on my phone :-)
> 
> What is being talked about here isn't really a second pass, but a
> true second pass is undesirable because it means that you have to
> wait for the complete page (well over a minute on many commercial
> sites) before you can display anything, or have to change the page
> after it has been displayed.
> 
> What's really being talked about is having a second set of combining
> rules that are applied to the inherited values to get the rendered
> values.
> 
> > Why can't I style a div to align centre using align:center but instead have
> > to use margin-left:auto; margin-right:auto?
> 
> Because this would interact with margins, and the result would probably
> be an even more complex specification (i.e. more rules for a competent
> author to have to learn).  It would be more complex because not only
> would you have to describe the results of various margin combinations,
> but also the results of combinations of your new property and margins.
> 
> Again, the point is that a standard has to consider all possibilities,
> not just what will go in the sales demo, so what may seem complex is
> possibly the simplest self-consistent solution.
> 
> Incidentally, the align attribute in non-strict HTML is a horrible
> combination of two different things.

Why not separate layout from presentation. Margins, borders and
padding aren't layout. They are used that way currently because no
other method exists.

I see margin, border and padding as the mechanims for how and object
uses the space allotted for it, not as a means of layout. If used for
layout, you get into all the nasty situations where you need to
calculated out, either by hand or the calc function, the widths of
things in order to make them line up.

Also layout isn't about flow, it's about space. Where things are
located on the screen or page. Layout doesn't have margins or padding.
It is simply boxes aligned in space with all of them flush together.

I make this distinction because the two use very different parts of
the brain. Finding things on a page uses the spatial part of the brain
and the brain learns the spatial orientation of things. Spacing comes
into effet when trying to read the space found after the spatial
orientation phase.

It's time that W3C styling specs took this into account.

Orion Adrian
Received on Thursday, 30 June 2005 04:33:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:38 GMT