XML and OpenSource

Key of XML initial deployment
Vast amounts of OpenSource and free XML software
Focuse on XML use within the community:
- Desktop project: Gnome, KDE
- OpenOffice: OpenSource Office Suite
- XFree86: core graphical environment
- Apache: server services
- Linux: the kernel
- Red Hat: provider of Linux and open source technology

SGML on the Web ?

Use of XML for Web content publishing

Not the most common use
HTML documentation generated from DocBook (KDE, Gnome)
Processing of XML format using PHP
Java is not commonly used (Apache, OpenOffice)

XML as a document format

OpenSource office software use XML as their native format

OpenOffice: switched to a ZIP container with XML content
KOffice: KDE office suite uses XML
Abiword and Gnumeric also use XML
etc...

Advantages in terms of format and I18N support are clear

Usually the files are compressed (ZIP or gzip)

Problem: they all use different formats, but ...

Storing application data

This is a trend, XML is used by default

"Why don't you use XML for ... file format ?"

Application configuration (bookmarks, drivers, templates, ...)
User's preference (GConf, XFree86 fonts, ...)
Any persistant data with textual mapping

The main toolkits used are expat libxml Qt XML Perl or Python

A Red Hat 9 system includes more than 3500 XML files

Storing graphics

SVG is making a hit !

It is getting the main format used for desktop icons

Endorsed by desktops environment (Gnome and KDE)
Scalable, high quality
Alpha channel composition
Processed as fast as rasterized formats

Gnome libraries librsvg, SVG editor sodipodi

Documentation == DocBook

The duck picture

Most projects have switched to DocBook

Program documentation
API documentation, manual or extracted from code
HOWTOs, help files

Use libxslt/xsltproc or Saxon for XSLT

FOP/xmlroff/PassiveTeX for PDF generation

Documentation Tools

Help browsers use HTML pregenerated or on-the-fly

Indexing and cross-referencing is ad-hoc

Various front-end are available:

ScrollKeeper handles documentation metadata
Per-project front-end: KDE Help Center, Nautilus, Yelp

WYSIWYG editing is not there (yet ?)

XML for User Interfaces

XML can be used to describe the set of widgets

Allows a modular design of the application

XUL used by the Mozilla project
Glade and libglade used by Gnome

Ease the development of the applications

Internationalization I18N support

Massive use of XML for documentation

I18N constraints, redundance, automation

Specific tools have been designed:

KDE xmlpo
Gnome intltool

Act as bridges between XML and the PO I18N system

XML for network use

Some use of SOAP (libsoup) and XML-RPC but not widespread

Jabber use is growing

Layers above XML

XInclude early adoption to add examples within DocBook

Schemas validation is limited:

Not that many OpenSource tools
It is complex
Maybe Relax-NG will get more acceptance

XSLT usage is increasing to convert to HTML or between formats

XML Catalog deployment

OpenSource and XML shared concepts

The appeal of XML:

Openness:
- self describing
- exposing the content
API at the data level
- just informal
- or backed up by DTD/Schemas
Reuse existing parser code
The XML format can be reused

Pieces missing and uncertainties

FAQ#1: "What editor for XML/DocBook are available ?"

More standardization at the XML format level: OpenOffice ?

Will Java and C# get deployed there, to which extend ?

Licencing and Patent issues
OpenSource framework
Reuse of native libraries is key

Conclusions

Use of XML is very widespead in OpenSource projects

Documentation and data formats are the main target

Use of core specification, and mostly C/C++ toolkits