Post edited 8:43 am – December 24, 2010 by wspademan

I'm hoping to have a sit-down talk some time next week with Van Mittal-Henkle. He is leading development of the Mifos software, and has responded graciously to my request to speak with him about Mifos and how it could be used for CGB. Mifos is well known for being developed by Grameen Foundation for microfinance institutions (MFIs). Since it is open source and we admire Grameen, and because they seem to have talented people working on it, and some funding, they are naturally worth looking at. If anyone would like to offer advice about talking points I should have for this meeting, please feel free.

Van says:

Mifos is currently evolving from an initial monolithic legacy app to a modular app. We're leaning towards moving to OSGi but no final decision has been taken.

The vision we have for Mifos as a truly modular platform would likely support the kinds of things you would need to do. The catch is that we're not there yet, so the timing of things would be something to consider carefully.

Mifos is geared specifically towards microfinance and so there are aspects of it that differ from conventional banking applications (the group lending model for example). However, many microfinance institutions are moving to the so called "teller model" which is much closer to banking as we know it. Because of this, we are evolving Mifos to have good support for individual lending.

I see the purpose of this discussion as clarifying such questions as: what is the timeline for improvements in Mifos, how close will those improvements move the software towards meeting more of our (yet to be fully fleshed out) requirements, what else could be done to help the Mifos team, and how interested are they in making any adaptations specifically targetting the kinds of use cases that we are interested in with CGB.

Some things to keep in mind: Mifos, as it currently stands, is very much a custom-made solution for addressing the needs of microfinance institutions in developing countries. Think farmer X in village Y goes to office Z with a list of things he needs to buy to expand his operations. The software then addresses how to set rules for a workflow prescribing precisely what farmer X has to submit, what kinds of followup the loan officer needs to make, and how those things get filed for later review by higher-ups. Also, it's by no means modular: right now, it can only handle these very basic kinds of MFI-specific workflows, and with a very specific user interface that is pretty antiquated, by the Mifos team's own estimation (which I fully concur with after downloading and playing with the software myself).

Perhaps I should really be trying more to draw the Mifos folks out on their larger vision, as opposed to dealing with specifics which at the moment we are somewhat short on. We don't yet have highly detailed functional specs for precisely what all our software has to do. Maybe what is more important is, are they interested in seeing Mifos become something that might be used outside microfinance, and outside the developing world, and if so, how soon realistically would they foresee that happening, and what kinds of support or incentives would expedite that happening?

On the other hand, perhaps there are a few more specific things I should be asking Van, based on the data modelling Kurt has done.

Once again, any thoughts you have to offer are most appreciated.