Wikipedia:Wikipedia Signpost/2011-01-31/Technology report

Update on Wikimedia "offline" project
As mentioned in the latest WMF Engineering Update (see previous Signpost coverage), Foundation developers have been working on a project to improve the availability of "offline" versions of Wikimedia content. These versions would enable readers to access the collective information of Wikipedia and its sister projects even in areas with non-existent or poor internet coverage, including India, which was recently visited by some of the Tech team. This week developer Tomas Fincz posted an update on progress with the project (Wikimedia Techblog, adapted slightly): The first step in making Wikimedia content available offline is to select it. The Wikipedia Version 1.0 Editorial Team has been steadily releasing new versions of their beta Wikipedia collections, but technical limitations have hampered how quickly those can be finished. We're going to evaluate the team's tool set to see how to support them.

Once the content has been selected, it needs to be packaged into a standard file format. The openZim format is an actively developed format for offline Wikipedia content, and we want to facilitate its integration into our general architecture. Our first step is going to be the enhancement of the Collections extension to support openZim. This will be done by our partners from PediaPress, who have already started to work on it.

After selection and packaging, the last remaining piece is the application that allows readers to access the content... During the strategic planning process, one app emerged as a good candidate for the WMF to actively support: Kiwix. In order to support this work and to help make the application even easier to use, we'll be conducting a usability study on Kiwix, focused on search and browse, during the first quarter of 2011. Later this year, we'll be focusing on an easier update cycle using openZim as the underlying storage format. We hope 2011 will be full of exciting news about offline Wikimedia content.

In brief
Not all fixes may have gone live to WMF sites at the time of writing; some may not be scheduled to go live for many weeks.
 * For the first time in several years, email addresses entered on registration are going to be checked for syntactical validity. Previously, addresses were only required to be even theoretically correct when they were confirmed before use at a later date (wikitech-l mailing list).
 * In the last couple of weeks a new customised "report card" for progress on improving support for Indic languages was released and commented upon (Gerard Meijssen).
 * Developer Happy-Melon prepared an extension to do away with the hackish convert template (revision #81074).
 * There was a discussion about the advantages and disadvantage of using chequered squares image backgrounds to mark transparent areas on file description pages (wikitech-l mailing list).
 * A list attempting to condense information on what remaining problems there were with the Vector rollout from the summer of 2010 has been updated recently (Gerard Meijssen).
 * 416 bugs were marked as resolved this week, representing new efforts to close bug reports that had been fixed in earlier revisions.
 * Developer Ryan Lane posted an update on the OpenStackManager extension he is helping to write (see also last week's "Technology Report").
 * The MediaWiki software for preparing database dumps was briefly broken (bug #27016).
 * The wordmark in the www.wikimedia.org portal was updated to conform to the new logo specifications (the font is now Linux Libertine, rather than the previous Hoefler Text).