The G String

Two months ago, Indymedia London launched a new version of an alternative London-centered news website using Ruby on Rails and Debian/GNU Linux. After several months of indecision about what system to use for content management; whether to continue development of an existing Java-servlet application, to use Drupal or to develop something new, the London-based collective have arrived at a topical activist news website that has successfully been adopted by the activist community, and is fresher and richer in content than ever before.

The system, called Hyperactive, is a rails application that embodies the strongest recent trends in web development. It make extensive use of RSS to syndicate content, including videos and comments a la YouTube and is soon to include mapping from OpenStreetMap on its events section. It has also been adopted by a Danish IMC, and looks set to become more established in the wider network.

It is a triumph of the open-source software development model, based entirely on free-software, and developed without traditional planning constraints or limitations. It is particularly pleasing that Indymedia bring their own non-hierarchical ideals to the process; a group of individuals from different backgrounds and countries are contributing to the codebase, although the group have never all been together in the same space, nor is there an appointed manager or team leader.

I’m glad to say I have been making my first commits to the codebase recently, contributing a ‘featured groups’ section to the bottom of the homepage. Starting from an idea over beers in a South-London pub, the feature was coded on the spot and turned around in less than a couple of weeks. In essence, the available open-source technology, from operating systems to IDEs to version control make it possible for part-time energy to result in professional and fresh websites.

The project demonstrates quite clearly what is possible, when a small number of unpaid developers achieve functionality not far behind huge corporate websites. Behind them are even more volunteers contributing to moderation, testing and promotion.

Yet the problems of under-capitalisation and lack of resources are still present. The huge corporations, with their advertising revenue and share capital have the advantage of huge bandwidth and hardware redundancy, while many of the alternative websites rely on donations and a dedicated following. The dilemma as developers is how continue innovating with open-source software whenever it falls behind the latest products from the software giants; for the open-source developer, nurturing new projects by utilising and contributing to them can sometimes mean offering an uncompetitive product, and in the field of news and politics there are huge resources available to the mainstream media.

But for the sake of a few hours work here and there, when you can see the difference made by your efforts; when it is fun to work on an open-source project, its hard to see the open-source and voluntary sources of energy ever drying up.

If you are interested in contributing to the development of the project, have a look at escapegoats project website.

When working in ruby on rails 2.0.2 I keep my development database in the default sqlite3 but use mysql for the production database on a seperate test server. I found that after a while of use from my alpha-testers, the test server database for my web app had been filling up with useful, real-life data, that I wanted to use on my development environment. So I needed a way to transfer the data from mySQL (5.0.32) into sqlite3.

I exported the data from mySQL using the mysqldump utility. The mysqldump tool is normally used for backing up mySQL databases, but with a but of cunning can be used to export data into other formats. Firstly, since my sqlite3 development database already exists and can be recreated from the rake migration scripts, I didn’t need any of the SQL statements that modify or create the schema. You can tell mysqldump that with the –no-create-db and –no-create-info arguements to mysqldump:

mysqldump --no-create-db --no-create-info yourdatabase

Secondly, mysqldump uses its comma-seperated INSERT feature for brevity. However, sqlite doesn’t like this, so I wanted it to use a sepereate INSERT statement for each row. This is done with the –extended-insert=0 option

mysqldump --no-create-db --no-create-info --extended-insert=0 yourdatabase

Lastly, I wanted to ignore any of the comments or locking statements and output only those INSERT statements into a file. I did this using grep and piping the output:

mysqldump --no-create-db --no-create-info --extended-insert=0 yourdatabase | grep 'INSERT'

you can then redirect this output into a file

mysqldump --no-create-db --no-create-info --extended-insert=0 yourdatabase | grep 'INSERT' > yourdatabase.sql

The one problem that I found with this output was that sqlite3 handles escape characters differently from mySQL, so I had to manually replace any instances of \’ with ” using my text editor.

This yourdatabase.sql file then contains a load of INSERT statements. Since I didn’t need the data in my development database anymore, I decided that the easiest thing to do would be to just delete the db/development.sqlite3 database and use rake to rebuild the schema, to leave me a clean, empty database. This avoids problems such as clashing primary key entries.

rm db/development.sqlite3
rake db:create
rake db:migrate

Finally, to import my data into sqlite3 I used the .read command in the interactive command-line tool

.read yourdatabase.sql