Wikimedia Foundation Relationship

Goals of openZIM

We divide the goals of the project in two parts:

being an efficient storage format for hypertext data and other MIME types needed to enrich the hypertext data (websites)
being free and open, so everyone can use it and implement tools for it
being a standard format for offline websites (dumps), so offline readers and data files can be freely exchanged and combined at full compatibility

provide a free and open source implementation of a toolchain to create and read ZIM files (zimwriter, zimreader)
provide a library which can be used in any application to add ZIM support to it, whether read or write support (zimlib)
encourage DVD makers and other projects like OLPC to use ZIM as a standard file format
provide a central library of ZIM files with open content (GFDL or CC licenses) with download mirrors, using the experience gained by handling the german Wikipedia DVD downloads

openZIM provides:

full documentation of the ZIM file format, so everyone can implement the file format using his favourite toolchain and programming language
zimwriter, an application which creates ZIM files out of dumps stored within a database (MySQL, PostgreSQL, SQLite, ORACLE)
zimreader, an application which reads ZIM files and serves them via HTTP, a webserver which can run on a desktop computer and has a very small footprint

Then there are some more tools provided by other projects:

Kiwix, known from the Wikipedia DVD created by Linterweb in France, is a reader application with a graphical user interface, like a browser and a built in search engine
mwdumper, a patched version of the original dumper from MediaWiki, used to create dumps and put them into the database, prepared for the zimwriter

category handling
search index creation (fulltext index, "old Zeno way")
cleanups of the zimreader code, removing old hardcoded stuff like application title and skin

long-term (topics on the agenda for the next Developers Meeting in Autumn 2009)

development of an UID scheme to be able to handle several ZIM files in parallel by the readers and use interfile links (like images)
evaluation of search engines and indeces to decide whether to use the Zeno fulltext index or integrate another search engine like Lucene

give approval to the openZIM project / Wikimedia CH to use the Wikipedia logo and trademark on the Wikipedia DVD issued as a give-away for LinuxTag 2009
- is now problem as long as we don't use the Wikimedia trademarks along with other companies logos (eg. sponsors). Just ask and the approval should be given soon.
be a fallback in the case Wikimedia CH decides to stop sponsoring the openZIM project
- The Wikimedia Foundation offers financial support if needed. There is no big budget at the moment, but a few hundred USDs should be possible.
developers, developers, developers...
what is the status of the wikitrust integration into Wikimedia wikis
- Preperation is going on currently with Luca and WMF to be able to integrate wikitrust. First only categories and other namespaces will be activated to see if it works out, later all articles should have this feature. But there is no timeframe yet.

there are ideas that Wikimedia provides dumps in ZIM files directly on download.wikimedia.org
- Tomasz Finc is now working on it. cxxtools and libzim could be compiled successfully on the Ubuntu server cluster. An interface to MediaWiki is now the next step.