XSL-FO rounds out the trio of standards that make up XSL. The FO stands for Formatting Objects, which are containers of content that preserve structure and associate presentational information. XSLT prepares a document for formatting using XPath to disassemble it and produce an XSL-FO temporary file to drive a formatter.
Under a W3C charter, the XSL Working Group started to design a high- powered formatting language in 1998. XML was still new, but XSL was understood early on to be a factor in making it useful. The group split its efforts on two related technologies, a language for transformations (XSLT) and another for formatting (XSL-FO). XSLT, the first to become a recommendation in 1999, demonstrated itself to be generally useful even outside of publishing applications. XSL followed as a recommendation in 2001.
Cascading Style Sheets, a jewel in the crown of the W3C, had been around for a few years and was a strong influence on the development of XSL. Its simple but numerous property statements make it easy to learn. You will see that quite a few of these properties have been imported into XSL-FO. The CSS box model, elegant and powerful, is the basis for XSL-FO's area model. Mostly what has been added to XSL-FO are semantics for handling page layout and complex writing systems.
The principal advantages of XSL over CSS are:
XSL is not always preferable to CSS. You would choose CSS when you want to keep formatting simple and fast. Most web browsers have CSS processing built in, but none can do anything with XSL-FO. XSL-FO processing is also very resource intensive . You would probably not want to push the burden of formatting on a fickle and impatient user when it could be done ahead of time on your server.
An important goal of XSL was to make it declarative rather than procedural. A declarative language expresses a desired outcome with constraints and high-level concepts that make it easy to read and understand. It also tends to be highly portable since it remains on a conceptual plane. By contrast, a procedural language is rife with low-level clutter such as variable assignments, algebra, function calls, and so on. It is typically harder to read and requires someone with erudite skills to work with. For example, PostScript, which is nearly impossible to understand unless you are a hardcore programmer, is procedural.
The XSL Working Group has strived to make XSL as flexible as possible. It is device- agnostic , supporting a variety of media from synthesized speech to on-screen display. It is designed for high typographic quality. It is multilingual, with support for bidirectional and vertical scripts, most modern fonts such as OpenType, and the full set of Unicode characters .
At the time of this writing, I find that XSL is still rather rough around the edges. Early adopters will find a variety of tools available with varying degrees of conformance to the standard, but none offer the kind of control you have with a proprietary publishing system like FrameMaker or Quark. If you want a WYSIWYG environment where you can specify pagebreaks and see the document repaginate before your eyes, you will not get that with XSL.
The choice of whether to use XSL may depend on two things. First, how much control do you need while authoring the document? Graphic designers need a very high level of control with instant feedback as they shape and sculpt a document. XSL implementations out there now do not offer anything like that. Second, how much of your workflow is automated? If you have huge stacks of documents to process, as is the case with legal departments and doctors ' offices, XSL can help. It works behind the scenes with no human presence required. Just set up your stylesheet, write a batch processing program and let it go.