CSS2 - A Rod for our Backs

[Jason Harrop <jharrop@speedlegal.com>]

My contention was that John is saying we absoutely must design our XML 
so that CSS is able to style it perfectly.

So I asked John -  Can you help the TC to understand what it means by 
showing us 2 things:

1.  Give us an idea of how the XML would need to look to satisfy this 
hard requirement. 

2.  Demonstrate that given the XML in 1 above which you need to give CSS 
its best chance, CSS is indeed up to the task.  The CSS will 
definitively answer questions such as the following:
   (a)   Do CSS implementations do hanging indents eg the number 
indented 1 inch, with the text following it indented a uniform 2 inches 
from the margin. 
   (b)   What browsers and other tools will display this CSS properly, 
and how does it look when printed.

John was kind enough to attempt both the above.

What was the result?

Regarding how the XML needs to look -  details are in a separate post.

Putting aside for the moment the complex XML, how did it look with CSS 
applied?  

First point - It looked okay in Internet Explorer but not Mozilla.  We 
haven't demonstrated our CSS presentation will work properly across web 
browsers.   Indeed, we demonstrated that it doesn't! 

As John said:

>I don't want to create false expectations -- I am not doing the Mozilla or any
>other cross-browser representations, believing that the purpose of this exercise
>is to show that standard W3C CSS, Level 2, is adequate to the task.
>  
>
I say, what is the point of crafting our XML around CSS2, if CSS2 is not 
implemented adequately across web browsers?  

I would say that _if_ we are going to require that our XML be such that 
it render perfectly in CSS, then the practical requirement ought to be 
that it is demonstrable using either IE6 or Mozilla (the current 
release, say).  Even this practical requirement is setting the bar high, 
since people using older browsers will have problems viewing the CSS 
with 100% fidelity.

If we take this practical route, we still cause problems for ourselves, 
since there are a variety of things in CSS2 which would be useful, but 
aren't implemented.  See for example the CSS2 selectors which are still 
broken in IE6: see http://www.xs4all.nl/~ppk/css2tests/intro.html

It is enough that our XML would look good if those selectors worked?

In short, its not clear what constraints CSS places on our XML, nor how 
we test our XML meets those constraints.  Faced with these practical 
difficulties, why would we include presentation (particularly direct CSS 
presentation) within the scope of our work?

Second point - the hanging indents.

>Insofar as "hanging indents", now I understand what you are saying ... yes, it's
>easy to do with CSS - simply add "text-indent:  -1.45em;" to the ListItem
>selector. There is a better, more sure, way of doing it that mimics what I can
>do with my markup for article, section, clause, and subclauses, because each
>article, section, clause, and subclause is composed of a BlockCaption and a
>BlockBody -- in these cases, all one needs to do is set both as inline objects,
>with the height on BlockCaption set to 100%. I have tried this internally and it
>looks just fine. 
>
I still haven't seen the hanging indents work properly.  Not saying it 
can't be done, but I did spend another 30 minutes experimenting with 
Mozilla 1.2.1 this morning...

cheers,

Jason