#seminaarikannu

W3C's Stephen Pemberton talking about Dimensions of Openness. Distribution of content is changing hands, Marshall McLuhan in 1964 said that if classified ads will move elsewhere, newspapers will fold.

Printing press enabled the Enlightenment, but it took a hundred years for that to happen. With newer paradigm shifts, like web, this will happen faster.

In 14th century when the New College of Cambridge was built they planted oak trees for renovating it, they were used in 20th century. This sort of planning for the future doesn't happen today.

Examples of bad technologies for openness: proprietary media formats, flash, HTML not being extensible. Web 2,0 also, as it locks data in silos. Good technologies include XML, CSS, RDFa, net neutrality.

Example: klm.com uses browser detection that blocks browsers they don't support out. I thought these practices died in early 2000s? There are also sites that block you out if the screen size is too small.

Guiding users to a 'mdot' is OK, but they should be able to browse the full site with their mobile if they really want to.

Accessibility: think also about the eyesight we will have when we're old, build the culture of the web to support it now so that you don't have to educate every new generation of web designers in their 20s about it

CSS and XBL will enable really rich visualizations on top of relatively simple HTML. Example show of a page having a list of cities and local times, rendered as fancy clocks using XBL and SVG.

HTML5 is worrying because they're removing things instead of making the markup extensible. 'Not enough people are using this attribute'.

Xforms can also become useful, you define 'a selection of four options' instead of specifying 'radio buttons'. Oracle has demos of serving same ordering form on desktop browser, cell phone, voice menu on phone call and IM buddy.

Walled gardens have a problem with Metcalfe's law (value of network is proportional to the square of the number of nodes). One web, one email distribution system is good, fragmentation of IM networks is bad. Since the value of a Web 2.0 sites comes from user-generated content, they tend to lock their users in. There are no standards for migrating data between services.

Because there is no data portability, you're also screwed if the service you made content with goes out of business, crashes, or imposes some terms of use condition that locks you out

The right way to achieve data portability is that the site marks it up with RDFa, the extensible way of defining microformats. 'RDFa is CSS for meaning'

RDFa also helps search engines and aggregators to serve you better.

APIs and XML feeds are just poor substitutes to having the data right there inside the web pages.

RDFa may mean disintermediation through aggregation: you don't have to post your job ad to monster.coma. just post it on your own site and let the search engines find it.

When will there arrive a website engine that is as easy to set up as WordPress, but gives you a Facebook-like experience using semantic technologies and aggregation? (maybe I should make an installer for the stuff that runs my site)

I've loved that Cambridge example since i first heard about it!! Good reporting @bergie

Time to give my first talk of today: Location-aware applications with GeoClue

Congratulations!

Pretty good audience, both desktop and web people. I have to upload the slides a bit later. Also, got to make some bitter remarks about how location is handled in Maemo 5 :-)

Jim Sangwine on bringing 3D to the Web. Started with a short video of Minority Report: 'They still haven't figured out wireless data transfer in the future'. Examples of rotating a tree and traveling through seasons in 3D with Flash and PaperVision.

New input methods needed: demo of Flash virtual reality, showing and manipulating 3D images drawn on top of computer camera picture based on symbols on a paper held in hand.

Another demo of symbols on fingers used for minority report -like image manipulation. @W_I 's motion sensor fingers would work here too, but people probably rather put on gloves with symbols than put technology inside their fingers.

Use all hardware capabilities for user input, most mobile devices and laptops have already the following in them: camera. microphone, accelerometer, digital compass, touchscreen. Combination of those can provide amazing user interaction.

Demo of Layar, an Android virtual reality app where you can see prices of houses when you walk around and point your phone at them. Another example of using Wikipedia as augmented travel guide.

Wii brought new, physical interactions to gaming and that changed demographics of the sector. Now other console vendors are also exploring the possibilities with Project Natal and that Sony 'magic wand'. Can the same happen to web applications?

Demo of a fighting game with Natal, kicking, punching and dodging using your whole body instead of some joystick. So we're really not that far from the Minority Report situation, only the display technology is lagging behind

Web GL APIs by Mozilla and Google will bring familiar OpenGL to the web

A remark from the W3C keynote: 'Start thinking of the web of 10 years from now, web of 2 years from now has already been thought about and done'

Talked with Paul Rouget from Mozilla. Jolicloud is working on an extension that allows authorized websites to talk to D-Bus on the computer. On Maemo this would provide access to everything: contacts, calendar, SMS, media...

Pretty inspiring stuff. With all these new possibilities I think it will become too difficult for organization to handle the whole stack. It may split in areas of comptence like:

• Data providing and content management with semantics. RDFa will be important here, as will content repositories like Midgard and CouchDB
• Aggregation, collecting data from various semantic sources and providing qualified collections of it. Google's strong area. Look at technologies like PubSubHubbub and XMPP for real-time data aggregation
• User interaction. New input and output methods, WebGL, XForms, SVG, JavaScript...

Sebastian Kügler on freeing the web from the browser. Web is both content and services, but so far things are quite limited to old-style computer hardware: ties to specific screen resolutions, keyboard mouse.

...but in reality web is used with many different devices, each with different display and interaction capabilities. Don't assume everybody has a keyboard, or can hover with a pointer. Pixel densities can wary also, compare 800x480 N900 screen to 640x460 media center browser on a huge TV.

Qt could be used to create richer web-connected applications. You can combine traditional web widgets, multimedia and WebKit views. This is the area where KDE Plasma is going to. Data engines and Applet UIs are separated. Everything is themable and scriptable.

Akonadi provides an interface for emails, contacts, calendar, bookmarks, microblogging, ... "Raindrop for KDE"

Nepomuk provides the semantic layer on the data. Tagging and rating, full-text search, SPARQL, ...

My GeoClue slides are up!

Project Silk tries to shift content and services from "HTML, JS and Flash" to something that you can take with you, "to solve the shortcomings of a web browser" and "to make Web a first-class citizen on the desktop"

Demo of Silk querying videos of Amy Winehouse from YouTube and Vimeo. The plugins are scripted, for example the YouTube provider is written in Python so it can be modified easily if YouTube changes the APIs.

Another demo of a comic widget for the desktop. A web service provides list of comics available. LOLcats also available for the desktop!

Freedom Services: looking at how well a web service complies with free software ideals. Some areas to look at:

• Specification available?
• Free implementation available?
• Privacy
• Data integrity, being able to move data between services

• Well-defined interfaces (XML, JSON, RDFa)
• Full access to content
• Separated logic from presentation

Silk demos:

• Social Desktop applet: show your contacts from open collaborative services, and people around you
• Selkie site-specific browser (showing a project on Gitorious in this case). The browser can introspect additional toolbar items based on site content, and integrate with system services like the notification D-Bus service

Time to give my "Midgard2 - Content Repository for desktop and the web" talk

Frank Karlitschek on Social Desktop: "Create innovation that Microsoft and Apple cannot deliver"

Shared API for social information between services: profile, file downloads, messaging. Geolocation is integrated so you can discover people near you who are using the same software.

There is a Knowledge Base API so you can integrate questions and answers about the application right into the app itself.

You can also display the profiles of developers of an application in the app itself. Users can become fans of an app, and get some special content like notifications and events related to applications.

There also is a tool for uploading and distributing applications through OCS. This would be useful also for Maemo Extras Q&A process.

Open Collaboration Services, the API behind Social Desktop is also used in the Maemo Extras application by @danielwilms

Paul Rouget from Mozilla: "We're not really competing against Safari, IE and Chrome. The real competition is Flash and Silverlight". All modern browsers (IE is not one) support the Open Web. Flash and Silverlight are black boxes.

Modern web is cool: video, audio, offline, SVG, transitions, Workers, Canvas, 3D, cross-domain AJAX, mediaQueries, ... But IE is holding all these back.

Cool demos:

• Image processing
• Geolocation and cross-domain XmlHttpRequest
• Remote fonts and drag-and-drop
• Video and CSS transformations to it
• Foreign objects in SVG
• Video player done in SVG
• Adding transparency to videos with 42 lines of JS
• Tracking movement in video with JS
• Controlling elements in a page through movement in video
• CSS transitions also look very good, smooth switches between two different CSS rules. Will be in next version of Firefox
• Another demo of upcoming stuff: accelerometer support via JS: browsing through a gallery of photos by tilting the device
• Face detection from video. teach about faces and visualize them

Official program ended and it is beer time. Some good chat over kreteks with Chris Chabot from Google about social network portability, PubSubHubbub and Salmon. Have to draft a plan have to bring some of those to Qaiku

Speaker dinner in the old police station. Saw demo of YubiDrone, Henk Klöpping's software implementation of Yubikey for Android phones. Would be great to have on Maemo too

@bergie Wow -- Android supports USB keyboard?

@Ile no, YubiDrone was a Yubikey done completely in software as the algorithms are open

@bergie Yep, of couse. There is also simulators, for example: http://wiki.yubico.com/wiki/index.php...