Offline readers 2009 notes

See also the static content group and related pages.

Notes from meetings and discussions in the year 2009 about offline readers.

Wikimania 2009, August 27


At Buenos Aires.



  • some interesting ideas from the Qi Hardware NanoNote (in Berlin)
  • openZim and its mailing list. possible new wikireader mailing list.
  • working on finding ways to get WP into print, on usb keys and sd cards, on cds and dvds. Need an overview of what's up in each space. producing images, tools for others producing images, media, distribution -- all could be separate pieces.
  • update from the land of olpc - the infoslicer activity is new, offers some enjoyable offline editing (though no synch with online services or wikis atm).
  • update from iphone work (patrick collison) - he's published his file format. more bare bones article versions than openZIM? (without some formatting), pre stripping format.
    • difference ZIM / openASCII
      • ZIM does not use wiki syntax but HTML - building a 3rd party MW parser does never fully work and rendering needs additional ressources on the reader
      • ZIM stores all kind of MIME types - including images
      • ZIM supports layout pages for readers that want to use a publisher-provided skin (eg. webbased readers as zimreader)

Current View of the Pipeline

  1. article selection - MW
  2. rendering and dumping
  3. storage and indexing with openZIM - MS
  4. displaying
    • various reader applications:
      1. OLPC - SJ
      2. Kiwix - Emmanuel Engelhart
      3. zimreader - MS
    • hardware vendors interested
      1. talks going on with KW
      2. technical contacts with TF

Various notes


EM - thoughts on what is good shape : format / pre-rendered formatted pages for a display. dump generation is well tracked - working it now, generated weekly.

MS - user participation of openZim adopters needs to be improved. we have adopters but the feedback loop is minimal.

BV - if you're doing version selection, you might want weekly info , pull from the recent dump, or optionally fetch newer stuff from the web.

EM - what about generating zim files directly?

EM - it seems to me this is something the wmf should really do, tim starling has reactivated static dumps before...

TF - let's figure out what to do with static dumps. making it sustainable from operational side. what to do when it blows up. sync up with tstarling about long term commitments and sustaining operations.

EM - we're missing the reader tool. [kiwix, moulin, wikibrowse] - none are really great. not awful or great. cross-platform reader apps we can port to new platforms. not sure java is it;

BV - you could have a core lib for reading the format, wrap it in a lightweight client that does the ui. can do it with a native web widget... whether it is moz or ie or whatever...

MS - we have the ZIM library on one hand and reader/writer application on another. that's c++ and you should be able to stick the lib in your favorite reader to make it zim-aware.

SJ - pieces to combine. the title-search code (from Patrick Collison). provide a way to say "give me this subset" for offline reader files.

TF - we totally want this, for dumps as well, and maybe make it eventually far more customizable. that would be awesome.

What to do next

  • Focus the consortium discussion on a public mailing list & meta - everyone
    • using wikireader at as project independent public offline reader mailing list - SJ
  • Bring back and support html generation of Wikimedia projects - TF
  • openZIM Developers Conference Nov 20th - 22th - MS

See also