Wikislices

From OLPC
Revision as of 18:27, 22 February 2008 by Sj (talk | contribs) (..)
Jump to navigation Jump to search

In this context, a Wikislice is collection of articles pulled from Wikipedia for the WikiReader activity on the XO. The goal is to select from wikipedia well written, structured, and cited articles while excluding the rest. The entire english Wikipedia is very large and wouldn't fit on the XO. Nor are 1000+ articles on Pokemon characters important education materials for the developing world.

Questions

Universalism

The question of universal use of this content needs to be considered. Do we run this project under OLPC entirely? Or do we try to create logical bundles for anyone with a wikireader? What are our ideas that may differ from other Wikipedians?

Bundles

We are planning on shipping a general collection of material with the XO. Additional packages will be shipped with the XS School Server or available online. Collections could be added to a student's XO based on classroom assignments or simply a child's interest in a subject.

Health

Science

  • An update to the Biology bundle with fungi and protists is slowly underway. Ditto an update clarifying licenses of the images (all cc-by)
  • Something for a bug blitz would be most helpful, drawing on the above and related zipcodezoo and misha h's content.

Bundling scripts

Scripts that are currently used to make bundles:

PDF and single-document exports

  • wiki2pdf

HTML dumps

  • wikiwix export (being built) : takes in a list of wikiwix entries, outputs ?
  • wikiwix interface (being improved) : allow selection via firefox plugin of a set of articles for a collection

Summaries and weight-watchers

  • Summarize list : takes in a list of article titles, outputs a directory of one-paragraph html files with css. (by Zdenek, not published yet)
  • Compress images : take a set of pages and images, reduce images according to a slider
    no images (remove altogether) v. hotlink images (include original thumbnail, alt text when offline)
    include first {0-10} images on a page, with metadata
    thumbnail only v. include full image (but not extra large) v. include all image sizes (full screen and more-than-fullscreen where available)
    bonus: assume local resize tool v. store 3 images for large instances

Older notes

see also the talk page

A Wikislice is a collection of materials gathered from a public wiki and packaged into a reusable form. Common examples are topical wikislices from Wikipedia, resulting in books such as the "Solar system" wikijunior text and various wikireaders. Tools used to make wikislices are regular expression toolkits.

See User:ZdenekBroz and the library grid for some examples.

WikiProject Wikislice

Please visit the project about wikislices on the english wikipedia.

Code libraries

  • KsanaForge and their KsanaWiki project have a set of scripts that process raw xml dumps from MediaWiki. They are working on producing read-only flash drives and SD cards for distribution.
  • Linterweb, developer of one of the freely-available static selections of Wikipedia, has an open source toolchain for building it; they are also working on wiki search engines (see Kiwix) and have offered to help build the local-filesystem search for the journal.
  • The Moulinwiki project and Renaud Gaudin have a toolchain from processing html output from the MediaWiki parser. They are now combining forces with Linterweb.
  • PediaPress has an "mwlib" library for parsing mediawiki text which is freely available
  • the "Wikipedia 1.0" team and Andrew Cates (user:BozMo on en:wp) is using their own scripts to generate and review static collections from a list of constantly changing wiki articles.
This article is a stub. You can help the OLPC project by expanding it.