OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Future of XSL-FO at W3C??

[ Lists Home | Date Index | Thread Index ]

On Friday 18 October 2002 10:53 am, David Rosenborg wrote:

> I find the current solution bizzare. I think developers are used
> to the pattern: XSLT for structure, CSS for decoration. 

Decoration? That sounds a pretty derogatory way of saying 'presentation' or 
'design'. It's much like going to an art exhibition and congratulating the 
artist on his choice of canvas, and oh, by the way, quite like the way he's 
decorated it with paint. Presentation is the important aspect because it's 
the phase that gets consumed by the reader - the end user, so from a variety 
of raw material it becomes possible to arrange and design editorial in ways 
that are conducive (or not) for consumption. Arguably it's far more important 
than the actual editorial, which is partly why we have style sheets in the 
first place - editorial matter will come and go, but if it is funneled 
through consistent and stable style sheets and templates, it will maintain a 
strong and viable identity. Such an identity is more important to a title 
than any individual issue or article.

> One way to do that
> would be to develop a standard XML representation of CSS. This
> would just be an alternative syntax just like there is an alternative

That's one of the interesting problems with XSL-FO - it's already way too 
close to CSS2. The property set for each formatting object is basically a 
choice from a large pool of properties, almost all of which are derived 
directly from - or at least without too much distance between - the existing 
CSS2 world. This, coupled with the quite lop-sided and inadequate supply of 
opportunities for actual page layout, belies the origin as simply one of 
trying to print web pages out in a formal manner, rather than adopting the 
clean-slate direction of designing proper publications from managed and 
composite dynamic sources. 

The XSL-FO 1.0 situation has strengths and weaknesses - the master-page 
pagination aspects are relatively strong. The layout - what you can actually 
put on a page and where - is astoundingly and perversely restricted and 
underdeveloped. The flow notion is quite odd, and the emphasis given to 
relatively useless items such as footnotes is again quite arcane. Floats are 
a very wrong way of doing what floats are trying to do, and if XSL-FO wasn't 
simply a reformulation of CSS2 in XML. 

Look at the internal document object model of (admittedly a proprietary 
solution, but one that works in the real world) QuarkXpress, which is simply 
a tool that allows a designer to place boxes on a page and get PostScript out 
the other end. Now QuarkXpress is designed for manual labour rather than 
automation, but the change needed to implement a XSL-FO-like structure in 
QuarkXpress would be less impactive and disruptive and involve less wheel 
re-inventing than XSL-FO seems to be suffering from. 

Not that I'm anti-XSL-FO - I'm not. I just happen to think that one of the 
barriers to it's growth is the lack of understanding surrounding it, and the 
odd lopsidednesses within it. It's as if it were a solution designed by 
programmers, to a problem space that designers could not express to said 
programmers, so they took the nearest thing they did understand and ran with 
that instead.

Ian Tindale


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS