Practical XML

Goals

• Improvement over HTML?
   
• Something to present with?

XML Overview

“Extensible Markup Language (XML) is a simple, very flexible text format derived from SGML. Originally designed to meet the challenges of large-scale electronic publishing, XML is also playing an increasingly important role in the exchange of a wide variety of data on the Web and elsewhere.” — W3C XML Introduction.

XML Syntax

• Syntax details.
   
• Stricter than (web browser) HTML.
   
• Closing tags mandatory.
   
• No overlapping elements.
   
• Quote attribute values.

Leading Elements

• Unicode Byte Order Mark (BOM) - optional.
   
• Some parsers choke on it.
   
• XML Declaration - optional.

<?xml version="1.0" encoding="iso-8859-2"?>

• Hint content is XML.
   
• Specify document character set.
   
• Default: UTF-8.

Validity

• Well-formed vs. valid.
   
• Document Type Declaration (DTD) - optional.

<!DOCTYPE eolas SYSTEM "http://sial.org/_xs/eolas/DTD">

• Identifies format (DocBook, SVG).
   
• Specification of allowed elements and attributes.
   
• External entity definitions.
   
• Alternative: XML Schema.
   
• Less well supported.
   
• Allows validation of non-XML content.

Processing Instructions

• optional hints to applications.

<?xml-stylesheet href="html.xsl" type="text/xsl" title="default"?>

• xml-stylesheet: associates transformation rules with document.

Namespaces

• Mix XML formats.
   
• Group related elements.
   
• Support common things.
   
• Really Simple Syndication (RSS).
   
• XML Linking Language (XLink).

<eolas xmlns:xlink="http://www.w3.org/1999/xlink">
…
<acronym xlink:href="http://www.w3.org/TR/xlink/">
<abbr>XLink</abbr>
<name>XML Linking Language</name>
</acronym>
…
</eolas>

XML Formats

• DocBook.
   
• Scalable Vector Graphics (SVG).
   
• Extensible HyperText Markup Language (XHTML).
   
• Extensible Stylesheet Language for Transformations (XSLT).
   
• XSL-FO.
   
• & etc.

Format Design

• Data vs. narrative formats.
   
• Data: CVS Change Logs.
   
• Narrative: howto docs, presentations.
   
• (Web browser) HTML poor starting point.
   
• Reuse existing formats, namespaces.
   
• Only where practical.
   
• Alternate namespaces complicate stylesheets.

XSLT

“XSLT is designed for use as part of XSL, which is a stylesheet language for XML. In addition to XSLT, XSL includes an XML vocabulary for specifying formatting. XSL specifies the styling of an XML document by using XSLT to describe how the document is transformed into another XML document that uses the formatting vocabulary.”

“XSLT is also designed to be used independently of XSL. However, XSLT is not intended as a completely general-purpose XML transformation language. Rather it is designed primarily for the kinds of transformations that are needed when XSLT is used as part of XSL.” — W3C XSLT Abstract.

XSLT Overview

• Taken off independent of XSL.
   
• Many implementations.
   
• Generally fast.
   
• Good for structural changes.
   
• XML thru XSLT to HTML plus CSS
   
• XML elements become CSS classes.
   
• Limited abilities on content.
   
• Future: EXSLT, XSLT2.

Work Flow

Discussion

• XSLT best producing HTML/XML.
   
• Well-formed HTML by default.
   
• Easy to produce varied XML output.
   
• Limited non-<> output capabilities.
   
• Poor whitespace, wrapping control.
   
• Little power over non-XML content.

Websites using XML

• XML instead of HTML files.
   
• XSLT renders XML to HTML.
   
• AxKit as glue.

XML Format

• Custom format for narrative documents.
   
• eolas.
   
• Simple text blocks, lists.
   
• Good for unix documentation.
   
• Alternative: Simplified DocBook.

Advantages over HTML

• Tags with meaning.

<a href="http://www.faqs.org/rfcs/rfc1869.html">RFC 1896</a>

<rfc>1869</rfc>

• [RFC 1869].
   
• More eolas examples.
   
• Separation of data from presentation.
   
• Customize on the fly.
   
• No pains to support lynx.
   
• Fancy site and WCAG/Section 508 support.

Stylesheets

• XSLT for each browser or output style.
   
• <xsl:import> minimize redundancy.
   
• Translate XML to HTML plus CSS.
   
• CSS for main stylesheet (eolas, DocBook).
   
• Tie data-driven output to existing CSS.
   
• Stylesheet choice via Processing Instruction.

<?xml version="1.0"?>
<?xml-stylesheet href="default.xsl" type="text/xsl" title="default"?>
<?xml-stylesheet href="lynx.xsl" type="text/xsl" title="lynx" alternate="yes"?>

Stylesheet Setup

• DocBook stylesheets source.
   
• DocBook XSL Stylesheets.
   
• Can be tricky to setup.
   
• Use local DTDs.
   

<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.0//EN"
"http://www.oasis-open.org/docbook/xml/4.0/docbookx.dtd">

• Faster rendering.
   
• Network outage tolerant.

AxKit

“Apache AxKit is an XML Application Server for Apache. It provides on-the-fly conversion from XML to any format, such as HTML, WAP or text using either W3C standard techniques, or flexible custom code. AxKit also uses a built-in Perl interpreter to provide some amazingly powerful techniques for XML transformation.” — AxKit.org Introduction.

AxKit Needs

• Apache 1.3.
   
• mod_perl.
   
• XML, XSLT Libraries.
   
• expat.
   
• libxml2, libxslt.
   
• Various CPAN Modules.

As of AxKit version 1.6; see the AxKit mailing list for discussion of potential problems with specific releases (e.g. XML and XSLT libraries).

Needs for AxKit

• Mapping browsers to stylesheets.
   
• Mapping XML to stylesheets.
   
• XSLT transformations.
   
• Static output caching.

Request Overview

Browser Requests

• Stylesheets customize content to browser.
   
• Via HTTP User-Agent header.
   
• Or manually by URL query ?style=.
   
• XHTML or HTML, as required.
   
• <xsl:import> reduce stylesheet redundancy.
   
• Cache for subsequent requests.

$ du -sh htdocs ../tmp/axkit-cache
2.3M htdocs
6.3M ../tmp/axkit-cache

Apache Config for AxKit

• AxKit documentation for details.
   
• Global setup in conf/httpd.conf.

<IfModule mod_perl.c>
PerlModule AxKit

PerlFreshRestart On
</IfModule>

Configuration Sections

Inside <Files>, <Location>, <Directory>, or .htaccess.

# limit what filetypes AxKit deals with
AddHandler axkit .xml .dkb

# use libxml2/libxslt for XSLT work
AxAddStyleMap text/xsl Apache::AxKit::Language::LibXSLT

# custom page instead of "500 server in flames"
AxErrorStylesheet text/xsl /_xs/axkit/error.xsl

# good for static content
AxCacheDir /tmp/axkit-cache
AxGzipOutput On

# make various request data available to stylesheets
AxAddPlugin Apache::AxKit::Plugin::AddXSLParams::Request
PerlSetVar AxAddXSLParamGroups "Request-Common"

Files & Exceptions

# file extension mapping
<Files *.dkb>
AxAddProcessor text/xsl /_xs/docbook/html/docbook.xsl
</Files>

# spare /wc3/p3p.xml from AxKit
<Location /w3c/>
AxResetStyleMap
AxResetProcessors
</Location>

Browser/XML Mapping

• Stylesheet choice in Apache config.

# map HTTP User-Agent by regex to stylesheet title
AxAddPlugin Apache::AxKit::StyleChooser::UserAgent
PerlSetVar AxUAStyleMap "lynx => Lynx,\
mozmac => ^Mozilla\/.+?Macintosh.+?Gecko"

# map ?style=lynx arguments to stylesheet title
AxAddPlugin Apache::AxKit::StyleChooser::QueryString

# map stylesheet titles to various XML formats
<AxStyleName "#default">
AxAddRootProcessor text/xsl /_xs/eolas/default.xsl eolas

AxAddRootProcessor text/xsl /_xs/docbook/xhtml/docbook.xsl book

AxAddRootProcessor text/xsl /_xs/cvs2cl/default.xsl changelog
</AxStyleName>

<AxStyleName "lynx">
AxAddRootProcessor text/xsl /_xs/eolas/lynx.xsl eolas

AxAddRootProcessor text/xsl /_xs/docbook/html/docbook.xsl book

AxAddRootProcessor text/xsl /_xs/cvs2cl/lynx.xsl changelog
</AxStyleName>

<AxStyleName "mozmac">
…

Random Tips

• Imperial character set mandates?

AddDefaultCharset utf-8

• Support for SVG in conf/mime.types.

image/svg+xml svg svgz

Alternatives

• Other XML handlers: Resin, etc.
   
• Caching: mod_proxy or squid.
   
• More flexible.
   
• Better for dynamic content?
   
• Pre-render XML.
   
• Easier to setup and run.
   
• Fallback site if AxKit down.

Conclusion

• HTML/XML same domain already.
   
• Separation of content from style.
   
• Content customized to user’s needs.
   
• Ease of new agent support.

Presentations using XML

• Custom XML format.
   
• XSLT renders XSL-FO.
   
• XSL-FO to output media.
   
• Slides for screen display (PDF).
   
• Handouts for you (PDF).

XML Format

• eolas format again.
   
• Single format to learn.
   
• Easy web display of talks.
   
• Alternative: DocBook Slides.

XSLT to XSL-FO

• No shareability with website stylesheets.
   
• Generates Formatting Objects.
   
• XSL-FO: “Making XML Look Good in Print.”

XSL

• Young format.
   
• XML, XPath, XSLT, CSS basis.
   
• 1.0 specification limitations.
   
• Slow.
   
• … ?
   
• Documentation for engine implementors.
   
• New book: XSL-FO.
   
• Engine-specific PDF Bookmark generation.

FOP

“FOP (Formatting Objects Processor) is the world's first print formatter driven by XSL formatting objects and the world's first output independent formatter. It is a Java application that reads a formatting object tree and then renders the resulting pages to a specified output. Output formats currently supported are PDF, PCL, PS, SVG, XML (area tree representation), Print, AWT, MIF and TXT. The primary output target is PDF.” — FOP Introduction.

FOP Needs

• Java 1.2 and up.
   
• Various Java libraries.
   
• Most in FOP distribution.
   
• JIMI Software Development Kit.

Usage

Need to render XSL-FO to PDF for presentation.

• xsltproc to render XML to XSL-FO.
   
• FOP Xalan XSLT engine weak, so use libxml2 utilities.
   
• xmllint to tidy XSL-FO for debugging.
   
• FO to PDF with FOP.

$ xsltproc fo-slides.xsl practical-xml.xml > practical-xml.fo
$ java org.apache.fop.apps.Fop practical-xml.fo practical-xml.pdf

• Makefile handy.

Limitations of FOP

• Young tool.
   
• Extra font support tricky.
   
• Poor error reporting.
   
• Java backtraces annoying.
   
• Sometimes cryptic.

$ java org.apache.fop.apps.Fop \
-xsl fo-slides.xsl \
-xml practical-xml.xml test.pdf
[INFO] FOP 0.20.4
[ERROR] null

XSL for Presentations

• Not a good fit.
   
• Presentations driven by style.
   
• XSL-FO better for data-oriented needs.

Publication Alternatives

• Other XSL-FO engines: RenderX, etc.
   
• PassiveTeX - XSL-FO to LaTeX.
   
• Roll your own.
   
• Parse and output via PDF::Create.
   
• HTML to published format (poor match).
   
• Links, other content lost.

Conclusion

• Multiple output formats nice.
   
• HTML, PDF, etc.
   
• Easy to target new media.
   
• XSL-FO so-so for presentations.
   
• Investigate Synchronized Multimedia Integration Language (SMIL)?

XML Usage Tips

• XML-aware editors: Quanta, XML Spy.
   
• XML well-formed testing helps.

$ xmllint --noout test.xsl

• Let tools format XML for you.

$ xmllint --encode utf8 --format raw.xml > test.xsl

• Utilities to assist content markup.

Working with XSLT

• Maintenance similar to shell scripts.
   
• Complexity proportional to DTD flexibility.
   
• Optional elements, attributes.
   
• Empty elements.
   
• Context of element.
   
• Keep limitations of XSLT in mind.

Testing

• Create test XML documents.
   
• Stylesheet compliance.
   
• New web browser support.
   
• Example eolas test document.

Creation

• New stylesheets will have startup costs.
   
• Add templates, functionality as format evolves.
   
• Catch unexpected elements in new formats.

<xsl:template match="*">
<xsl:message>
<xsl:text>Unmatched: </xsl:text>
<xsl:value-of select="name(.)"/>
</xsl:message>

<xsl:apply-templates/>

</xsl:template>

• See unexpected-test.xsl for complete example.

The author disclaims all copyrights and releases this document into the public domain.

Questions or comments about this page? Current ruminations available on my blog.

$Id: practical-xml.xml,v 1.1 2004/01/22 19:27:42 jmates Exp $