Archives/Technical track

From Wikimania

Daniel F. Abawi "Using Wiki Systems as Information Back-Ends for e-Learning applications"

Motivation - Problem Analysis & Requirements - Concepts & Architecture - Use Case: The LernBar E-Learning Environment (@ Universitaet Frankfurt am Main)

Why integrate a Wiki with another system?

Wiki will not replace ex cathedra classes More & more professors at Frankfort are asking for Wikis to be setup. Allows for more detail on topics.

Example Implementations at JW Goethe U. Frankfurt: Computer Graphics (CGPedia), Japanology (JWiki), Medicine (AnatomyWiki), Theology (Basis-ReliPaed)

Integrate -- Use of existing specialized or legacy applications to present course content or manage classes. Many professors believe that a Wiki system improves student interaction with the class. So why not replace existing system with Wiki? Risk of loss of data.

Still want to use the Wiki as a web application; however want the applications to talk to the backend containing the data. This presents problems: 4 issues: applicatoin has its own form, needing the data to be transformed; differences between the Wikis & their implimentatoin; visual & navigational characteristics of diff. Wikis; online/offline access.

4 tier system to allow structured query of Wiki data.

Multi-tiered integration architecture Wiki abstraction layer/Wiki layer - example in Java or p-code, asking to return info related to "eyelid" article, article itself may then return related articles. - example of GenericWikiWrapper using objects to make use of polymorphism.

Use-case involving LernBar. Courses broken up into a number of web pages. On-the-fly integration of the glossary, thus if a student encounters a term they do not understand, they can consult the glossary thru hyperlnk. (Visual example showing pop-up info about human eyelid from the Wiki.

Need to take care to use the Wiki as a Wiki, & not abuse it as a content of data. (Talks as if this this has not be implimented yet - or at least beyond an experimental basis.)

Q about the display of the data as RO. Design decision, forcing users to modify the info on the Wiki itself.

Q: Is prototype real or a mockup? Real.

Q: is this a screen-scraper? No, takes the data from the Wiki database. But depends on how the wrapper is written & applied. Also about API, about lack of structure in the Wiki data. Datastore for off-line access.

Q: Worry about multiple implimentation of Wikis, keep the number small & concern over collapse of a Wiki.

Sponagl: Device independent access to all Wikipedia projects

Showing an experimental application. Mobile phone access to Internet. More popular in Japan (e.g.) than thru computer.

The challenge of creating a website that can be easily viewed on a mobile device. Showing the application on a real Handy (Nokia 3650) shown over a webcam; (buttons being pressed by a robot -- to the fascination of the audience).

Using the existing browser in the Nokia, user will get the message "Seite zu gross" However going to "en.wikipedia.7val.com" -- shows a modified page that can be displayed.

Q: Is this using an algorithm, or was it hand-done? A: Can do both algorithm or "tweak" it.

Showed tab-completion with app. Can even edit with the cell-phone.

Quote Jimbo about everyone should have access to Wikipedia. States this is a technical issue for the next few years. Typical access mechanisims: web tablets, appliances, TVs in home, mobile devices (phones, PDA) 3 of these devices to each computer with a ethernet connection. Emerging countries need mobile access to knowledge; thus we need to support Playstations, Nintendos. "Transform Playstations to knowledge stations". "Access to Wikipedia anywhere, anytime, anyhow."

Reducing content so that it will fit on the limited viewing field of a mobile device. Service & content adaption: Wikimedia software functionality often is a superset to what is needed. The browser on every Handy has its own set of bugs,

Sevenval FIT is a transcoding & service oriented proxy. Number of sensitive customers using it, including banks because it is very secure, does not touch the data. Written a mobile configuration for WP. Supports 2000 mobile devices & all Wikimedia Projects. wikipedia.7val.com has been up & running since Jan. 2006; 15,000 page views a day without any marketing, 20% per month growth.

Sevenval motivation: marketing value, feature feedback -- & fun. "Supporting wikipedia is fun. I use Wikipedia every morning" Wikipedia motivation: New users, extends reach, prevents feature creep.

Q: What is your business model? A: Wikipedia just for marketing, have other paid customers e.g. tmobile vision. Four days to develop the Wikipedia application.

Q: Asking about use of this on his cell phone, Google's PDA version, a pda.wikipedia.org.

Q asking to define screen-scraping.

Q: Are people using this to edit Wikipedia? A: Don't know. Will show the IP of the val7 proxy.

Kasper Souren: Nokia 770

DICT protocol (RFC 2229), used this to be able to put Wikipedia on his laptop. Not a well-known format, not much support on Windows but comes bundled in Debian.

Wrtoe wik2dict in Python.First version took a day to process. Working verson up online. Not his server, but found someone using his script to convert. Shows example page "search" which preserves hyperlinks in article. However the version of WP he converted is an older version (example article is a stub).

Q about how it runs. Needs to run a dictionary server.

Describes the Nokia 770. Then started on MaemoDict. Started writing the client, but as of April 2006 MaemoDict includes a dictd server. Uses the Eagle library. Working to add WikiTravel support. FBReader support, "one of the best e-book readers"

Can use wik2dict to extract specific articles.