If you want to read the first post in Sand in the Sandwiches, click here
After Mauritania, I did one more trip for NRI as an employee, and two further trips as a consultant for them. Then I moved to the Virgin Islands. During that trip to Mauritania, I snapped. Something inside me told me that short term consultancies like this were really not the answer, that we needed a bigger gameplan both for our own sanity and to ensure we kept our clients up with the progress. I argued it out with both Bob and Judith one night. I lost the argument, but only because Bob really was a client and Judith reacted to clients wishes and would do anything for anybody. My problem was that neither of us were really computer experts, we were geographers who used computers to solve our problems. I think we did it rather well and got a reputation in NRI as fixers of GIS. But we never thought like computer programmers, and if you are writing complex applications which involve data, tools and interfaces, you need to plan very carefully how what the client wants before the work was done. Although we had Ould Babah over a few months before, it was really Ba we should have talked to, and when we arrived in Mauritania with what we thought was a finished product, we found that we had to constantly make revisions and updates. Some of these were trivial, spelling changes or new colour schemes, others were far from simple, although it was often hard to explain to the clients why. I spent long hours during the day while Judith trained updating programs, reworking the database structures, and that work spilled over into the late afternoons and evenings back at the apartments. When the changes were made, there would be further problems. Occasionally they would change their minds and I would go back to previous versions (one thing I had learnt was never throw away the older versions until you were sure you had finished). At the end of the whole trip, we still had a huge wishlist from Ba of what other changes he wanted in the database, the problem being that we had now used up the money for the project and any new changes would be under goodwill, something Judith was willing to give, but myself, with work in the Caribbean and elsewhere to get ready for, I just did not have the time or energy to do things for free.
So my argument with Bob and Judith was that the whole project was topsy turvy and we developed the system incrementally instead of planning, and I said we had to draw the line somewhere and say no to some of these changes. Bob in particular saw that as a refusal for a computer expert to meet the needs of the client, which happens to be my biggest complaint of software writers who would rather develop something they know instead of answering the clients needs. My work on the RAMSES, Lake Tanganyika and Caribbean CRIS had shown I took to heart making systems useful for other people. I was quite hurt when I heard it being used against me. I could not get them to realise that what I was complaining about was the whole process of developing such an application; we should have iterations, there should be a prototype that is fully tested by the people who have to use the software day in day out, not the bosses, we then have a full version and then a final version and at each stage we needed to interact face to face with the clients, no matter how far away they were. I vowed there and then that I would never work in this rather enthusiastic but haphazard way again, and I never have.
Mauritania was bloody hard work. Most of my trips were bloody hard work, but even here, where I got very little time either to myself or out of the rigid routine, I managed to have some wonderful experiences, see marvellous sights, and above all, despite the fact I seemed to be turning into a tired old traveller, there were some incredible novelty to come to grips with in this enchanting country.