W3C home > Mailing lists > Public > www-style@w3.org > October 2003

Re: CSS 4?

From: Werner Donné <werner.donne@re.be>
Date: Fri, 24 Oct 2003 11:09:02 +0200
Message-ID: <3F98EC2E.10909@re.be>
To: David Latapie <julian27@ifrance.com>
Cc: www-style@w3.org

David Latapie wrote:
> On Thursday 23 October 2003, at 23:57PM, Werner Donné wrote:
>> I don't see why both wouldn't continue to exist and be supported, despite
>> their overlap. XSL-FO describes only a layout, while CSS describes how
>> style should be applied to a document, which is a combination of two 
>> things.
> I never understood the "layout" term. Not just a language problem, but 
> also about the concept. Layout appears to me to be a "sum of style", so 
> with no added value compared to CSS styles.
>> XSL-FO is more oriented towards paged media. The spec provides a lot of
>> support for paging. CSS hardly does, at least for the moment. In the 
>> printing
>> area this has created two cultures. Browsers, the most common CSS UAs, 
>> are
>> very bad at printing. They can't even cut pages properly. XSL-FO tools
>> are mostly very good at it. CSS itself, however, does not precludeine
>> printing.
> What about CSS/Print? Granted, they may have to be fine-tuned (first of 
> all, browsers should support CSS/Print, later we may consider improving 
> CSS/Print themselves).
> As for me, "paged media" = "print" (as well as projection) and "print" = 
> "CSS/Print"

I have said CSS doesn't preclude fine printing. It is a matter of culture.
Browsers could already print decently today, but it simply doesn't seem to
be important to the browser implementors.

> I may have miss a point

No, you are right. CSS/Print is indeed a good direction to go.

>> The two are perhaps growing towards each other. CSS3 provides more paging
>> support, while using XSL-FO interactively in a browser is just a 
>> matter of
>> mind set. Conceptually, browsers could provide as much functionality with
>> plain XSL-FO files than they can now with HTML. XSL-FO has in fact the
>> potential to be more performant in a browser than XHTML+CSS, because the
>> decision of what should applied where and in which circumstances does not
>> have to be taken anymore. The server has already done this.
> Let me see:
> "Joe_browser: hello you! please serve me page1.html
> -Bill_server: OK, who are you?
> Joe_browser: I'm Mozilla 2.1, no picture, no JavaScript, no Flash. I 
> need blind-supprt as well
> -Bill_server: Alright. Let me compile this... here we go"
> Did I inderstood the principle of XSL:FO? A sort of custom-made 
> preprocessing? A template/master-document?

It is not a sort of custom-made preprocessing. In XSL-FO you describe a complete
formatting structure. It is not necessary to produce it on the fly. A web-site
could be full of XSL-FO files instead of HTML files. They would be produced with
a content management environment at publication time. Only for dynamic content
XSL-FO would have to be generated on the fly.

Note also that there isn't much difference when it comes to including pictures
scripts, etc. XSL-FO has the required elements to host all that, so it could be
exploited by UAs.

For dynamic and static content XSL-FO has the advantage that data can be served on
the web without even using HTML. This makes it easier to use other XML vocabularies.
If you send XML+CSS to a browser today, say Mozilla, you have a nice but dead
document. If you would instead produce XSL-FO, the same XML could result in an
active document. The driver to generate this XSL-FO could even be CSS, because
it is convenient to express style with.

Let's take the notion of a hyperlink for example. Today you can only use HTML
or XHTML to tell a common browser that some area is a link. I think this should
not depend on the authoring vocabulary, HTML in this case. There should be
generic way to express this. XSL-FO is such a way. There is, of course, also the
possibility to use XLink. But then the XML vocabularies must have schema's to
allow mixing in of foreign namespaces. This would require an amendment of most of
today's XML vocabularies. Going via XSL-FO is much less intrusive.

>> CSS and XSL-FO can also work together. The latter can be the 
>> representation
>> of the layout defined by applying the former to a document. In other 
>> words,
>> you can convert XHTML+CSS into XSL-FO.
> For instance, for user-designed style sheet

Werner Donné  --  Re BVBA
Engelbeekstraat 8
B-3300 Tienen
tel: (+32) 486 425803	e-mail: werner.donne@re.be
Received on Friday, 24 October 2003 05:12:51 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:09 UTC