Wikipedia on iPhone

This is one of the most missed features of my Palm and Windows Mobile smart-phones. The ability of having the entire world knowledge in the palm of your hand can prove to be a great boredom killer and a great asset sometimes. Most importantly, OFFLINE.

This was something that the iPhone cannot do, until you have jail broken the iPhone. The package that i used is called Wiki2Touch. The details of this project can be found in http://code.google.com/p/wiki2touch/. Note that you might require the MONO project (http://www.mono-project.com) to be installed if you want to be able to build the image indexes on your own.

Specific things to note are the instructions on how to index a database and then collect the images. The entire thing is a extremely lengthy procedure. Especially the downloading of the images. Downloading and the initial indexing of the articles (no images) took me about 2 days to get done. However, the downloading of images took me about 2-3 weeks.

Searching online doesn’t reveal anyone that has the latest articles and images. So, i have decided to put up my article only index online for those who are interested to use it. The file is divided into 2 parts. After downloading, you have to join them back together before you can gunzip it.

cat file1 file2 > finalfile

You can download it from the slower main site here.

The file size is pretty big, 3.2GB to be exact, and the bandwidth can be rather slow. Be patient. Also, copying the files over wifi into the iPhone can also take a few hours. Be Prepared!

Compressing all the images into a bz2 format resulted in a image pack of almost 8GB in size. Compare this to the main article index which is already 3.2GB in size, putting this into the iPhone will result in problems loading in your own music and movies!

You can download it from the slower main site here.

If you want to be able to use these files, i would expect you to read the wiki2touch pages and understand how to install these files. The instructions on how to create the image packs is available at this link.

NOTE: the pack command takes the following known arguments

-c:            build no destination file but count only
-o filename:   use 'filename' for the output (default is 'images.bin')
-s max_size:   if an image is larger it will be skipped

It would be good to note that including the images for Wiki2touch WILL result in ALOT of space being taken up. So, how would you decide on how much images to include? i have done some analysis and here are some charts that can possibly help you in making the decision.

From the above graphs, you should be able to decide how much space you will use up when determining the size of images to truncate. in the above index, the maximum file size is 8.96GB! If you are interested in the excel spreadsheet, it is provided in the link here.

Some other interesting site to check out on Wikipedia statistics includes

NOTE: The articles file and the images is based on the wikipedia dumps updated on 2008-10-18.

NOTE: you will also need gzip or bzip2 to decompress the download files before you can use them.