XTech 2007 was an interesting conference, lacking in Wi-Fi but made up for by the speakers. Here are some notes (slides are often available from individual talk pages):
AMEE - Gavin Stark
One of the opening keynotes at XTech, on the launch day of AMEE (the Avoiding Mass Extinction Engine), a web service for calculating energy usage, produced in cooperation with the UK government and climate change groups.
Check if government is going to supply appliance energy consumption measuring devices.
Networked electricity meters, consumption displayed in real-time as devices are switched off.
Heat map of energy consumption at residential addresses (by major postcode area to preserve privacy).]
Joost - Antoine Quint
Joost uses Mozilla's XULRunner to overlay a compound document (Gecko UI) of XHTML, SVG (scalable graphics) and XBL (widgets) on top of a full-screen video background.
Uses SPARQL queries of RDF metadata store to provide 'saved search' channels.
Had to introduce custom sprite elements, rendered as OpenGL textures, to break up the DOM into smaller pieces for performance reasons.
Waiting for Gecko support for DOM 3 events, SVG interactivity + animation, synchronous DOM events.
RDFa - Elias Torrez
Use rel, class, property and about attributes to add structured markup to visible data.
Originally designed for XHTML 2.
Can use in valid XHTML 1 if use the right DOCTYPE (modularisation).
Can use in HTML 4, but don't mind lack of validation.
May be able to use in HTML 5.
Give namespace as xmlns attribute in containing element, along with about attribute (URI for the subject).
Rel (and class?) attributes contain namespaced content, the names of which can be concatenated to the namespaces to define the type of data.
Single parser for all RDFa. Now supported in Operator, so can use alongside microformats.
CSS3 for printing - Hakon Wium Lie (Opera), who was also showing off an OLPC $100 laptop
Currently use Prince to convert XHTML to PDF for printing.
New features for CSS to provide the harder print features.
@page for margin elements, such as headers and footers
image-fit: contain for image scaling
string-set for variables
footnotes pulling in content from elements selected by id (this was ugly, to deal with legacy content)
web fonts (provide urls for fonts)
book microformat, based on DocBook; uses class names.
Nabaztag - Rafi Haladjian (Violet)
An amusing talk on Nabaztag(/tag) and its rabbit factor: the aims of the project, observations on the development method and ways people use such a networked object.
Document-centric XML authoring - Tetsuya Tashiro
A tool from JustSystems for editing compound XML documents. After defining a view and a controller for a particular XML schema, the XML can be transformed for viewing and editing; changes made to the document are then linked back to the original source, which is updated.
Geospatial data, a two-way street - Eric Schuyler, Tom Carden
People contribute lots of geodata.
Tools for organising and tidying up those contributions.
XForms, REST, XQuery + skimming - Mark Birbeck
XML fragments from separate source + XForms => dynamic form UI with widgets appropriate for datatypes (in future Firefox, currently only in plugin for IE). GET and POST data over WebDAV, from eXist. Can also use Atom and GData. Needs cross-domain requests (security settings in Firefox 3, hopefully).
Google Base - Jeffrey Scudder
Store arbitrary metadata, no limits to amount stored; shows up in searches (structured or otherwise).
Google Data API - Frank Mantek
Data model uses extensions to Atom (can be overly verbose for some applications, but can live with it). APIs uses Atom Publishing Protocol; interface to lots of Google apps.
Auth tokens for third party authentication.
Client libraries for many languages, but easy to implement interface to REST anyway.
Attributes and suggested values chosen usen ranking based on popularity.
Augmented Wiki for Scientific Visualisation - Frank Marchese