bq. If I understand correctly your position, you are advocating CSS over XSL-FO here because CSS syntax is easier and you are assigning properties directly to elements from source XML document. I think that I can agree with your position here”¦ but only as long you are using CSS with some general, document oriented XML format like XHTML or DocBook.
Yes, this is an important part of the argument. CSS is well suited for structured document formats where the content comes roughly in the order of presentation. I believe content should be in this near-presentation state when it “crosses the wire”. Styling should be applied as close to the reader as possible, i.e. in the client.
The other argument for using CSS in printing is that one can reuse many of the CSS style sheets written for the web.
bq. This example clearly shows that there are classes of documents which must be transformed to some more general markup prior assignment of visual characteristics.
I agree completely. And CSS hasn’t been designed for that purpose. XSLT has, and is perfectly fine to use. It’s Turing-complete and can perform the computations needed to calculate your columns. My only problem with XSLT is that it has “Style” in its name.
bq. You told me previously, that ToC for your book was created with some script.
Yes, we use Bert Bos’ “multitoc”:http://www.w3.org/Tools/HTML-XML-utils/ to generate a TOC. There have been proposals for how to handle this in CSS, but it’s probably too much of a transformation thing to make it into the CSS standards.
bq. XSL-FO is a such intermediate markup. I can imagine that you can also use XHTML+CSS for this purpose. But you are loosing big advantage of CSS then—your CSS rules are no more working against original markup, but against intermediate XHTML code.
I don’t see any problem with working against ‘intermediate code’. I think the XHTML code is what you should offer on the web since it uses well-known semantics. Your invoice example uses tag names not universally known. That’s fine as an internal format, but shouldn’t be published on the web. Also, I “don’t think XSL-FO should be published on the web”:http://people.opera.com/howcome/1999/foch.html—but that’s a different debate :-)