It is often helpful to think out loud on my blog, where I can be more
informal and less structured. =) I'm working on an actor-network
analysis of open source in developing countries, so I'll think about
pieces of it over here before editing it into a more scholarly form.

Okay. Network dynamics.

It is instructive to start with the closed-source view of the world.
Software developers in developing countries can take proprietary
software solutions such as Microsoft Office or the Oracle database
server, develop solutions on top of it, and sell these solutions to
the local market. This allows developers to meet the needs of the
market without spending a lot of time writing everything from scratch.
The solutions also gain credibility through their association with
global brands. However, this presents certain problems:

Cost. Although studies of the total cost of ownership show that
labour costs far outweigh license costs, these studies do not reflect
the case in developing countries where labour costs are lower and
licence fees are disproportionately high due to weak currencies and
other factors. See really crazy chart of GDP per capita vs licencing
costs for Microsoft Office.

Trade imbalance. Think of all those dollars flowing out of the
country… I heard that Microsoft partners make 9 dollars for every
dollar Microsoft makes – but that just means that 10% is going out of
your economy, versus open source which lets you keep all the value-add
inside the country.

Lack of deep access. Without access to source code, developers
can't customize closed source programs to really fit local markets
through localization, customization, integration, etc. in ways
unanticipated by the global developers or in ways that were not
profitable for the global developers to support.

Dependence. Local software developers become dependent on the
proprietary software companies, which could change its licencing terms
or discontinue product lines.

Random Emacs symbol: gnus-article-strip-all-blank-lines – Command: Strip all blank lines.