Re: XSL-FO versus PostScript

Could someone explain the advantages of XSL-FO over PostScript? As far as
I can see, XSLT can be used to output both,

Certainly. However, the FO processor does quite an amount
of computing, for example justifying spaces, adjusting lines
if there are different font sizes used and perhaps hyphenation.

and PostScript looks like the
more powerful language. Can't I exert more control over my documents if
I go directly from XML to PostScript

Certainly. However, get an OS XSLFO processor and look what
it does. If you want to use XSLT to directly generate PS,
you are restricted to fairly easy and most of all to a
stable layout, because the more complicated cases (like
footnotes in changing page masters or aligning page numbers
depending whether they are onn odd or even pages) get
soon out of control.

The one argument I can see in favor of FO is that it presents a method of
page description that is more intuitive than PostScript. But even that is
kind of iffy. XSL-FO goes through a lot of hoops (like inheritance) to avoid
programmatic flow control. And the result is fairly convoluted.

XSLFO was the first really systematic attempt on formalizing
higher level layout. Of course it contains some cruft, mostly
inherited from CSS (the most notable immediate predecessor).

If I'd put in charge of redesigning it without concerns for
compatiblity, I'd change a few bits.