Re: [xsl] Prince XML vs Docbook

["Eliot Kimber ekimber@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

I've been using XSLT+CSS to render XHTML to PDF using Antenna House Formatter.

I presented on my experience with using CSS for pagination at the DITA Europe conference: https://www.slideshare.net/drmacro/using-css-paging-to-render-dita-documents

tl;dr:

The cost advantages of CSS are so compelling that they outweigh the limitations in layout functionality (which can be compensated for either through source preprocessing or the use of formatter-specific extensions) and the current poor state of CSS pagination how-to information (basically the specs are impenetrable and there is no CSS pagination book or even web site that I've found).

As Liam says, CSS is not as complete as XSL-FO in terms of its layout features. In addition, the specs are still under development and many important aspects are underspecified. 

However, CSS is so much easier to work with and is so much more accepted that the cost in functionality and spec fuzziness is far outweighed by the ability to use less-specialized personnel to do the styling work.

Basically, if I never have to generate XSL-FO again I will be very happy (and I was one of the first power users of XSL-FO and a contributor to the spec).

CSS+XSLT provides a better separation of concerns between the styling concern and the data transformation concern. The result is that the XSLT required to prepare the source content for styling is much simpler than the equivalent XSLT-ot-FO transform would be and it is easier to create and maintain the CSS. For example, a web designer familiar with CSS for browser delivery could easily learn the addition pagination stuff and then do most or all of the pagination design using their normal development tools. The main challenge I ran into was the fact that browsers don't implement or even necessarily recognize pagination features in CSS, so you can't preview everything in a browser, but you can certainly verify all the content formatting aspects. You still need to use your PDF renderer to test and validate the pagination aspects.

Coordination is still required between the XSLT and the CSS but it can be at the level of hierarchical structures, tag names, and class names, rather than low-level details of how the FO should be structured.

Another potential advantage is that the same CSS can be shared between PDF and browser delivery, making it easier to create and maintain consistent presentation for publications delivered in both PDF and HTML to browsers. At least in the DITA world today, using the main open-source tool (DITA Open Toolkit), the PDF tool chain and styling definition is completely separate from the HTML generation and styling, making it more expensive than it needs to be to keep the two styles in sync.

For my project (which is non-DITA) and for most DITA projects it will be necessary to use extensions to the base CSS functionality. Antenna House Formatter definitely has these (basically everything it can do for XSL-FO it can for CSS through one kind of extension or another). Prince and Vivliostyle also have similar extensions.

For my project, which involves producing "change pages" I had to also do XSLT post-processing of the AHF-generated area tree, which is a literal XML representation of the formatted pages. This allowed me to implement a multi-pass process that gives subsequent passes access to details about the pagination of the source (for example, I capture the mapping from source elements with IDs to the pages they produce). I was also able to do tricks with IDs and empty elements in order to inject information into the area tree needed to support my point-page post processing. I would have had to do something similar with an XSL-FO-based solution, so here the use of CSS is not really germain, just pointing out that when your rendering tools support it you have the opportunity to do a lot.

Cheers,

 

E.

--

Eliot Kimber

http://contrext.com

 

 

From: "Vasu Chakkera vasucv@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Reply-To: <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
Date: Wednesday, January 17, 2018 at 6:30 PM
To: <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
Subject: [xsl] Prince XML vs Docbook

 

I came across a tool called Prince XML , which apparently works on Jason/XML and converts it to PDF or HTML or any format using CSS .

I did not work on the publishing area for a while, but the last time I did, we were working with docbook , DITA, topics and applying XSLTs on the XMLS to generate XSL:FO documents for PDFs and XSLTs for HTMLS .. are there any good arguments for one or the other ?

Are people still using the good old docbook XSLTs ? Any pros and cons for the Prince XML vs XSL:FO? Or am I comparing two totally different things?

 

Vasu Chakkera

EasyUnsubscribe (by email)