Re: [xsl] XSL-FO versus PostScript

Subject: Re: [xsl] XSL-FO versus PostScript
From: "W. Eliot Kimber" <eliot@xxxxxxxxxx>
Date: Fri, 28 Feb 2003 13:10:37 -0800
David Carlisle wrote:

Having said that, while I am a supporter of FO, given that most of my
documents are English and/or mathematics (ie no vertical or
bi-directional text no fancy bordered effects) I would always
(at the moment) print XML via XSLT stylesheet to latex rather than to FO. (see the pdf versions of the MathML spec for example)
It's hard to know in my case though whether that's based on
a real unbiased assesment or just a product of having latex
somewhere burried deep into my subconcious after having worked on it for
15 years or so.

Modulo limitations in the XSL 1.0 spec, a good FO implementation should be able to produce results comparable with TeX (and of course, there are two TeX-based FO implementations, Sebastian's and Epic 4.3, which is TeX-based underneath). For MathML, there's no reason in theory that one couldn't integrate a MathML-to-PostScript process as part of a larger FO process, possibly even leveraging existing TeX-based MathML processing.

Obviously, TeX has been around a long time and is very mature. FO implementations are much younger and therefore are less complete and tuned, but there's no *inherent* limitation in the ability of FO-based systems to produce typographic results as good as those produced by any other system. All the main FO implementations are being aggressively developed and refined by their developers.

This has probably been covered already, but it's also important to point out that PostScript, by itself, provides no facilities for doing typesetting--that is entirely handled by whatever it is that generates the PostScript. PostScript is a graphics language--it's up to a higher-level system to use that language to place marks on pages appropriately. To that degree it's not even meaningful to compare XSL-FO (or TeX) and PostScript.

The value of XSL-FO accrues from its generality and abstraction, which allows developers of presentations to leverage the hard work of FO implementators, who in turn can leverage the hard work of typesetting engine implementors (if they use an existing typesetting system like TeX). This value includes ease of development and maintenance, as well as vendor independence, at the cost of some limitations in functionality. If you can live within these limitations then the value of FO is very high from a business standpoint. If you can't, then you have no choice but to look for other technological solutions.

For individuals implementating essentially a one-off solution to a problem it might make more sense to craft a TeX-based solution, for example, or even to craft direct generation of PostScript, especially if you have more time than money.

For enterprises, which tend to have more money than time, concerns like cost of implementation, cost of ongoing maintenance, limiting project risks, and protection from proprietary lock-in and dependency become much more important criteria, and within that equation, FO-based solutions look very attractive if the rendition requirements can be met. It can be the difference between a 10K Euro solution and 100K or 500K Euro solution to the same problem.


W. Eliot Kimber, eliot@xxxxxxxxxx
Consultant, ISOGEN International

1016 La Posada Dr., Suite 240
Austin, TX  78752 Phone: 512.656.4139

XSL-List info and archive:

Current Thread