Opinions and motivations of a tech-skeptical software developer @ an NGO working in education
Building a future with equal access to learning opportunities (via technology!) is a long and winding road: We need to consider sustainable scenarios for our technology and face up to true challenges — not be tempted by hype or simple fixes.
About a decade ago, when bandwidth was a sparsity in rich and poor countries alike, designing and developing offline-friendly technologies was prioritized. Software ran on computers without internet access, things got mirrored on local networks etc. Not least, we were used to sharing documents and media with CDRs and USB flash, a.k.a. The Sneakernet.
Almost any computer or laptop could be opened with the same screw driver.
From my experience of working with schools in Malawi since 2011, offline education technology support has been an increasing challenge. Open access contents, educational resources and information databases have become more and more dependent on online platforms. One example is Khan Academy. It’s a highly complex and interactive website (with a simple UI!), only available to people with internet access. But back in 2011, it was a static website that could easily be downloaded and made portable on a hard drive and shared on a local network. Khan Academy provides great content, and they allow other actors such as Learning Equality to extend this content into the offline domain through a permissive license. That’s how KA Lite was started (and eventually how I came to know about Learning Equality and join the team to help build platforms for sharing educational resources offline).
Technicians (in Mzuzu, Malawi) installing software, replacing faulty hardware, instructing teachers: they are truly enablers, because they create the necessary environment for all the changes that advocates of a digital education would like to see. We can talk about technological enablers, but it’s actually people who maintain it, explain it, and configure it, who are key. (Photo: Benjamin Bach, CC BY 2.0)
Back in 2011, I was working with FAIR Denmark, and we left 4 refurbished rack servers running real Wikipedia (not just a static copy) on Ubuntu+Apache+Mediawiki+MySQL. They were installed in secondary school centres, serving approximately 50 computers through a cabled Ethernet. It had taken months to download and make Wikipedia work, with its 600+ extensions and ridiculous amount of image files. Finally, people could see photos in their original resolution and complete a free-text search in less than 10 seconds. It was actually pretty cool, because it was the same experience as having a real Internet connection: The slow search was compensated by the fast local bandwidth. We thought it was fine, but lo we forgot to calculate how fast Wikipedia was growing. 2 years later, it was evident that images could no longer fit on a hard drive, and the database was too big for text searches on a simple desktop server.
In 2016, I visited a centre that had originally had its server installed in 2011. Mponela and Malawi in general are dusty and warm places, and the hard drive with the database and all the images was an external USB 2.0 drive, consumer-grade. Surprisingly after 5 years, the rack server still worked, and the Wikipedia contents from 2011 were still showing (full disclosure: after 6 years, it went to lunch). Classroom computers that hadn’t developed hardware faults were still booting Edubuntu 10.04. Other places had maintenance and hardware swaps, but this particular server was the real deal. Stable software and stable hardware ❤.
The consideration here is not the encounter of an abnormally well-functional set of hardware and software, but real evidence of something that’s possible: Imagine what can be done educationally when an ICT centre Just Works™! Imagine a software and hardware platform that just spews out educational contents, and does so for years and years without any need for software updates, adapting to new functions or user interfaces. Just uninterrupted and simple delivery of educational contents.
2011: Refurbished rack servers were available in abundance, and there was a low complexity of distributing online through local networks. So we did that, but online resources have gotten more complicated since then. Also: Almost any laptop and computer could be opened only with one type of screw driver. (Photo: Benjamin Bach, CC BY 2.0)
Lots has changed since then: It’s not possible to scrape huge websites like Wikipedia without having a customized platform. Ubuntu no longer fits on a 700 MB CDR. Application memory footprints are increasing (you need at the very least 1 GB to run a modern desktop), and user interfaces are changing faster than educational books are published. Developers are shifting focus from the desktop applications over to web platforms, so for instance we no longer have Edubuntu, the education-flavored version of Ubuntu.
Most experiences and software setups deployed some 6 years ago are now obsolete. But I have to say: Back then, things were more stable, changed less frequently, and were less complex in their nature.
The world back then kind of didn’t need Learning Equality, but now it certainly does!
I’m skeptic about most things that relate to technology, but working for Learning Equality is not a compromise. I’ve seen many dusty computers providing little or zero value for recipients in development projects, and I have also seen digital content sitting stale on servers because of a lack of training or understanding of relevance (see also Shivi’s post about Too-Muchness). I know that KA Lite is not some magic quick-fix for education: Amongst many decisive factors, it needs teachers, and those teachers need training; It needs computers, and those computers need electricity.
By allowing ourselves to be honest about these challenges and listening to experience, we can navigate our way to address the root causes of poverty and inequality, by removing barriers to quality education caused by dependency on broadband internet, intellectual property and technological disruption.
Recognizing the threat of complexity and disruption, is especially important for technologists who design and develop software that should have a long-term scope, the life-cycle of which does not include daily automated updates.
January 11th marked 5 years since we lost Aaron Swartz, an activist and leader of making information available to those who need it the most. To some, he broke the law, but to me, he sacrificed everything he had for a better world. Sharing information with someone who needs it is a moral imperative, as he put it in the Open Access Manifesto.
I always loved open source as an obvious choice for humanity: Why can’t everyone just share their software inventions freely? If competition is about everyone sitting in their own corner, re-inventing their own wheels, I’m out.
As a side-effect, I caught an allergy for acts of redundant inventing/innovation/disruption hype. I think by now, half my work is about support, community, integrating and helping other people’s contributions, and staying in touch with other people in the field — the other half of my work is software development. And that’s a good thing, because imagine all the re-invented wheels, if we didn’t coordinate our work and sharing our visions, plans and technology across teams and organizations.
Making true open source is an attitude of collaboration: communication and integration with the outside world, not just some legal license.
In order to create true value, we need to highlight the labor of maintenance, and to not get obsessed with innovation. It’s about long-term commitment to keep software functional for the coming generation of computers, and long enough for schools and teachers to learn about the platform, collect experience and integrate it; repeat. These processes are not likely to benefit from technological disruption.
Seeing open source as a principle in aid and education can help to reduce or even eliminate the old power divides between rich and poor countries. Not wanting something back and not retaining ownership is critical to allow for the freedom of a society and its educational institutions to develop.
My work could be said to be in the scope of ICT4D, Information and Communication Technology for Development, a very vibrant and intersecting field, both in academia, amongst NGOs and in the commercial aid sector. I would prefer to argue against this as a constructive term, although it may unify a lot of organizations, people, and projects! Technology is never neutral, and neither is the perception of development. They both quickly become terms that are too broad and overloaded, justifying criticism of manipulation, alternative interests, not addressing fundamental structures that uphold the global inequality, poverty divides etc. Those which we ultimately seek to eliminate as a consequence of our actions of global solidarity and compassion. I prefer to work with education, teacher training, tech support, maintenance and sustainability: The lesser hyped, but classical terms. Nonetheless accepting those true tools in our long struggle towards eliminating poverty and creating equal opportunities for our fellow humans.
We develop and share open source, and we contribute to other related open source projects. We are a balanced team of developers, designers, documentation authors, and education specialists.
Technology cannot solve this alone. But creating technology that helps to give access to education and information to those that need it the most is certainly addressing those root causes.
We are not building a prototype for this. We take all types of devices into account and work with a broad scope of education and deployment specialists, and we encourage anyone with an idea or a contribution to participate.
To grow our strength and potential, we are building our own unique organization, making lots of room for discussion, and applying horizontal organizational principles that joins everyone in strategy and visionary level. And I love this.
At this point, the Learning Equality development team has started to widen its focus by recognizing our ability to contribute upstream. This is especially true in the Python community, where we interact with lots of other projects, thanks to the affordance of Github.
We’re working towards integrating the offline platform Kolibri (our flagship project) into a range of projects focused on the delivery of educational resources offline (e.g., IIAB, IdeasBox, RACHEL, Kubo, Edulution). Those projects contribute back and help shape the platform in ways that we cannot internally anticipate and design for.
Not least, we are not the sole group of people or organizations: other great platforms such as Kiwix are also addressing offline access and distribution. By participating in their hackathons, we have built many new relations and expanded our horizon, as well as replacing the isolated instinct to compete with an love for collaboration. Together with these teams, we are coordinating protocols, data formats and interoperability.
For the future of Kolibri, I dream that we will manage to polish and engineer our technology all the way until it aligns with the very strict and highly regarded Debian policy and becomes an official open source package for Debian/Ubuntu/Raspbian/Mint etc. Not least, all the other Linux distributions. This way, we can adapt our technology to a finely polished and long-standing line of software that was & is sustainable. State-of-the-art.
We have hundreds of ideas, bugs and rough designs to sift through and prioritize in the coming months, and that makes me think about how this platform, which is open source, does not belong to Learning Equality, but belongs to everyone. We cannot realize all of the ideas alone. That’s why we built a plugin API, that’s why we write documentation, and that’s why we try to make everything as straight-forward and inviting for community to contribute.
If you need me, I’ll be on Github! Thanks for reading, feedback, anecdotes and likewise are most welcome in the comments section :)