Kiwix install at Kwena Malapo school Johannesberg

By Andy on Monday 1 March 2010, 12:47 - Computers - Permalink

Kwena Computer lab Kiwix, an offline wikipedia selection, is installed at Kwena Malapo school.

I have blogged before about creating offline copies of wikipedia for use in school computer labs that do not have internet access.

Wikimedia, the umbrella organisation behind wikipedia and other related projects, is also keenly interested in offline uses of wikipedia. From November through January, the offline taskforce had a series of IRC meetings where we attempted to answer questions relating to the use of these offline copies. We came up with four recommendations, one of which addressed schools, and another cellphones.

In conjunction with this effort, Emmanuel Engelhart made progress and could deliver me both a compressed zim database of 30000 articles (2.4 Gigabytes, including pictures) and the software, called kiwix that could read and deliver this database served by http - the standard web protocol.

There is another tool, kiwix-index, that will read this large zim database and create an index. If this index is supplied to kiwix-serve then a floating javascript search box appears, and the whole article collection can be searched.

I was asked by Mark Batchelor of the Sandton Rotary club to help him with the computer lab they installed at Kwena Malapo secondary school. This school, with principal Michael Maligana, has 800 students and operates out of simple prefab buildings close to Lanseria airport in Johannesberg. The computer lab is a converted shipping container. A new door had been put in, with four windows and roof ventilation. A network of donated computers running Windows XP and Microsoft Office had been installed.

My proposal to Mark was to install kiwix as a 'poor mans internet' - an rich information resource navigable via a browser, with hot-linked words that drill down to more articles. We all recognise this as Internet, but the principles of the information age could now be brought to a school that has no internet connectivity.

Ricardo, Mark's Windows technician, and myself travelled out to the remote farm school on a saturday, and met Michael Maligana. I commandeered one of the computers, installed Ubuntu Linux, kiwix-serve and its software dependencies, the huge zim database and its associated index for search. We networked all the computers, and set the browser home page on each to the kiwix front page.

After some networking and power hassles and four hours in the hot environment we finally got it all working. Kwena had 30000 wikipedia articles selected especially for a school environment, and was one step closer to the information age.

Update:

I wrote a blog post describing how to get the software and install it.

Comments

1. On Monday 1 March 2010, 17:16 by Bronson

Hi Andy, that sounds like an awesome solution for creating an internet like environment for the kids, nice work.

2. On Monday 1 March 2010, 18:13 by Aragon

Very cool project, Wizzy, well done. I wonder if there's a way to keep the offline data updated without consuming much bandwidth? I'm thinking along the lines of rsync and dialup. Would be nice to sync it once a month or so!

Aragon, the 2.4Gig is a burden even for me. Even I would like rsync, or some way to unpack the zim file to a directory, rsync that, and re-pack it. Andy!

3. On Wednesday 3 March 2010, 01:17 by Kelson

@Aragon

There is currently no easy way to keep a ZIM update without downloading the whole new file. This is not an easy technical topic. We will try with openzim to propose a way in the future to do that.

People with a concrete project and unable to download a ZIM file may contact me.

Kelson, if you formalise things, perhaps with debs, tgzs, in a fixed download directory, I can arrange for it to be mirrored at mirror.ac.za. A useful rsync would still be very nice - http://svana.org/kleptog/rgzip.html might help ? is the zim a collection of compressed files, or a compressed collection ? Andy!

4. On Thursday 4 March 2010, 00:54 by Kelson

@Andy

I can already make easily debs (you too with "make deb") and also rpms

I'm working to provide asap a PPA and maybe other solutions for other distributions (have a look to the opensuse build service)

Source tgz will be published soon on sourceforge with the first 0.9 alpha. In the meantime I provide SVN tgz here: http://tmp.kiwix.org/src/nightly/

There is also an official FTP to mirror here ftp://download.kiwix.org/kiwix/ but this is only stable content (a bit outdated) and all ZIM files and 0.9 code are unstable so have to be kept on http://tmp.kiwix.org. An other issue is that the ZIM format not really stable is, we are working hard to improve it.

About the incremental update, ZIM is more complicated than a gzip file. But we want to make something analog: publish a "diff" zim file and provide a tool able to merge this "diff" zim file with the old zim file to update.

To answer you: a ZIM file is a collection of compressed collections + indexes by titles and urls + many other meta datas.

5. On Thursday 4 March 2010, 19:18 by Martin Walker

@aragon
One problem with updates is choosing a version that is free of vandalism - not easy. The collection Wizzy uses needed manual checks, which took several months. However, the WikiTrust people are working on an automated solution, and I think within 6 months - 1 year we MAY have a good method. If the ZIM file can be created easily too, this will allow regular updates in the future, but as it stands you'd have to download the whole file.

6. On Monday 7 June 2010, 08:23 by anthony

One problem with updates is choosing a version that is free of vandalism - not easy. The collection Wizzy uses needed manual checks, which took several months. However, the WikiTrust people are working on an automated solution, and I think within 6 months - 1 year we MAY have a good method. If the ZIM file can be created easily too, this will allow regular updates in the future, but as it stands you'd have to download the whole file.

Wizzy Africa