A long time back, when all of us were still thinking up magical things to do with Ankur, the most often talked about point was the price point of computing.
At LinuxAsia 2005, the last session on ICT4D hovered over and over again to the need to make computing pervasive by delivering services which are of immediate requirement. Through its various issues, i4D has also attempted to bridge the gap between localisation theory and localisation practice. The question that is relevant is that – are we there yet ?
If one looks back at how the computing hardware prices are moving, it is obvious that far better hardware platforms are available at appreciably lower prices. In this respect, Joel’s post is important.
A complement is a product that you usually buy together with another product… All else being equal, demand for a product increases when the prices of its complements decrease.
Step back for a moment and visualise the Operating System as a product. In a simplistic interpretation of the above statement, demand for OS should increase if the prices of hardware decreases. Or in more telling terms, prices of OS should be moving upwards in sync with the downward movement of computing hardware. Now, extrapolate the Operating System as a base for providing services. Thus, with the increasingly affordable prices of computing infrastructure, the demand for services would increase. This is exactly what is happening in the Indian FOSS space. So the lesson in here is that the more the complements are commoditized, the better chances the Application Software vendors have.
Does this sound like a model of Service Oriented Architecture gone wrong ? It could if you look at the wrong end of the spectrum. For a moment take a pencil and paper (or fire up the calculator program) and do some simple math. Take ‘x’ as the number of NGOs who are actively engaged in ICT4D domain (‘x’ < 'X', where 'X' denotes the total number of NGOs in the country), consider 'y' as the lowest common domain serviced by such NGOs. Multiply 'x' by 'y'. This gives you the total subset of application domains. Straightaway eliminate around 40% of the figure by assuming that these would in someway be linked to proprietary technologies. That brings a consideration of the remaining 60%. How many of such deployments would be common ? I would hazard a guess as to many-many and surprisingly enough, most would be doable through a LAMP stack. When GRIND was being considered, it was based on the concept of being a base Operating System which will lend itself well to be customised towards a L10n Services Framework. I think that one of the ways such a framework concept could grow is by following some sound project management. So are we there yet ? I think we are there at the initial stages. The concept is sound, a few deployments around it have produced desired user feedback. And what’s more, being based on the fundamental principles of Free/Libre Open Source Software Development, taking ahead a community driven participatory model would not be difficult.