Difference between revisions of "Yearly Report 2009"

Jump to navigation Jump to search
740 bytes added ,  11:36, 23 March 2010
Line 44: Line 44:


=== November: 2nd Developers Meeting, Major Changes to the ZIM Format ===
=== November: 2nd Developers Meeting, Major Changes to the ZIM Format ===
[[Developer_Meetings/2009-2]]
The second Developer Meeting again in Schopfheim, Germany, attracted further participants. Beside the project team itself the Wikimedia Foundation was represented by Tomasz Finc, the CEO and a developer of the french company Linterweb which works on Okawix was present as well as two persons from Qi Hardware, a company based in Berlin selling small embedded device like the Ben NanoNote.
The 2 major things about the meeting was the support of the [http://en.qi-hardware.com/wiki/Main_Page Ben NanoNote] platform and major changes in the file format.


2 guys from the qi-hardware team joined our meeting and we helped them to port zimlib to the Ben NanoNote.
In several discussion rounds the ZIM file format was further enhanced and clarified. Quite oppositional use cases had to be united into a new standard: Small devices such as the Ben NanoNote only can deal with a limited set of markup and need to strip everything but the article data from the format to save memory, while as Kiwix wants to add additional metadata for each ZIM file and article.


We decided also to change the file format to support indexing by url and title. Also compression was changed from bzip2 to lzma, since lzma is faster to decompress and library support is finally good enough. We also dropped the QUnicode format inherited from the zeno file format, since it only helps to sort german data better. So we decided to simplify ordering and use simple string compare.
With a special meta data namespace for ZIM meta data and a secondary article namespace for per-article meta data and templates these problems could be solved. A unified integer compression, a dynamic MIME type lookup table and a second article index based on URL cleaned up the format and solved some other long-standing issues.
 
In the end even the long-wanted lzma compression was tested and considered being finally usable. All these changes broke the compatibility with the old format but introduced many new future opportunities for the ZIM format.
 
For more details dee [[Developer Meetings/2009-2|http://openzim.org/Developer_Meetings/2009-2]]


== Budget ==
== Budget ==

Navigation menu