In FAIR Denmark, we've been learning by doing. One of the cornerstones in our project is to deliver cost-efficient educational ICT centres in countries where it's not affordable otherwise: And refurbishing ICT hardware creates not just a little more, but a lot more access to information and ICT education.
But does it also lead to development? This is another question, which is fundamental. At one end of the spectre, ICT can be seen as a tool for development, while at the other far end, it's seen as development itself. In either case, there's a firm belief that ICTs lead to development, thus we can dedicate our focus to our own experiences and how we can improve our practical delivery of ICTs that simply work and can be maintained in a cost-effective manner.
In February/March, we visited a number of centres set up in 2011. We wish to apply our new-found observations to keep elements in our model that work and replace the ones that don't.
Some of the routines that we are adjusting have to do with our collection and workshop activities at home, and some of them have to do with the abroad maintenance scheme.
Identifying equipment with a long life-span
The problem: An unacceptable fraction of the equipment from 2011 did not last 5 years. The explanation is of course different from computer to computer, and from screen to screen. The dust and the heat is harsher at a school in Malawi, but some equipment stands the test, and that's the kind of equipment we would ideally be sending.
The solution:
- Knowing how old equipment is: We will only be accepting equipment of at most 6-7 years old. This means that we expect equipment to have a total life-span of at least 12 years. Precise assertions of hardware age is hard, but once we have established and communicated this intention, we can make the distinction in our workshop for individual machines. If we are in doubt, we can reuse parts of the machines.
- Different grades of hardware: Much consumer hardware isn't built to last for long, however as our focus is with professional equipment from businesses and institutions, the majority of hardware has already been of a better quality.
- Stress testing: In the future, we should be able to relief the most obvious cases of over-heating, memory faults and hard drive failures by testing or reading out such hardware data.
But do we waste computers?
No, the success criteria is to identify hardware that won't run for another 5 years and not to falsely identify hardware that can!! Furthermore, equipment that's deemed too old is taken apart for spare parts. Hard drives, RAM, graphics cards -- much can easily be picked and shipped for later usage.
Maintenance, maintenance... maintenance
Maintenance falls into two categories: The day-to-day maintenance conducted by the schools, and the technical maintenance conducted by technicians appointed by the local project.
Problem: There have been obvious faults in the way that equipment has been cared for. Broken windows that leak dust (Malawi and many other African countries are really dusty), lack of curtains, lack of cleaning, and bad initial setups causing usage to strain power and network cables.
Furthermore, we also need more frequent technical maintenance. It's no use to keep a computer alive for 5 years, if it's out of use for half the time. If there's the more or less common problems of a broken hard drive or memory block, a technicians should be available and able to move quickly. This is not the fault of any individual person, but we see the need for technicians to be dedicated to the single task of visiting schools and for schools to have a clear demand for maintenance.
The solution: Before setting up centres, FAIR and partners are now much more aware of the initial communication with the schools. These efforts are mainly manifested in a MOU - Memorandum Of Understanding. In the first project, responsibility of technical maintenance was the schools' after 1½ years, but in the next projects, the schools will cooperate and finance a unit at a university to provide this, while the economical capacity to establish an individual body is being sought out.
This should make responsibilities more explicit and ensure that resources are there, not just temporarily, but as a permanent structure.
Keeping functional hardware up while it's still working
The problem: Even though hardware may not be working, software failures may lead to useless workstations. In the Windows world, this is usually caused by viruses, but in the Ubuntu / Linux world, it's caused by the missing local presence of a capable technician.
The solution: Efforts have already been done to ensure that re-installing a workstation is a matter of a few key presses: Boot a computer from the network, select re-install, and then with no further efforts, the operating system and all of its software is back.
But we need to be better at spotting recurring issues at the same work station. This will in the future go in a locally stored technical log.
Procedures and manuals
The problem: Every time a new Danish workshop volunteer or abroad technician start, we refer to personal training, and loosely collected manuals. This causes different practices, not an optimal procedure.
The solution: Refurbishing hardware at the Danish workshop should be picked up from start to finish to ensure an optimal recovery of reusable hardware. Lots of great efforts have been done, but it's the last fraction of un-vacuumed motherboards and over-aged machines that we need to guarantee does not occur.
At the schools, we need to make sure that technicians have a manual to solve problems and efficiently setup centres. This will both ensure an easier deployment of new centres, and should establish a cheap and swift routine of maintenance visits to all the school centres.
And yet again, why not new hardware??
Yes, the question arises often, and we feel tempted to criticize other ICT4D efforts for being too focused on bringing the newest hardware. So let's imagine that we had lots of money.
- Would computers last forever? Clearly not, they are bound to break and need maintenance.
- Would they last longer? Yes, if the equipment purchased is a kind of equipment that's robust enough -- we prefer solid desktop machines over tablets, whether new or refurbished, however we do acknowledge circumstances where light-weight and low-energy modern tablets are the only option.
- Would hardware be uniform at all centres? Clearly not, contemporary hardware of 2011 would be different from 2015, so equipment would still vary, and so would skills and spare parts needed for maintenance. Using a technology where spare parts are interchangeable between different models and generations is preferable..
- Would we need spare parts? Of course, if we want to fix broken stuff, we would still need something to fix it with. And where would we get that from? Another fictional pile of money or would we...... refurbish?
- Are we doing development if we spend 10*) times as much on new hardware as on refurbished equipment? Well, comparatively, we might as well stick to reusing hardware and use all the fictional money on solar panels or new school buildings.
*) “10 times as much” is our own conservative example figure. Buying new equipment is easily 10x the cost of shipping donated and refurbished equipment, provided that the refurbishment process can be carried out for free, and that collection and warehouse costs are kept reasonably low. - Wouldn't new hardware consume less energy? Not if you do the full calculation. Main parts of the energy footprint happen in production. Running your technological items for as long as possible is essential to minimizing CO2 footprint.
We are not just refurbishing because we have to or want to: In any scenario, reusing resources should be a first-choice.
But we can learn and improve.