Open education licenses are global, so the infrastructure that open resources run on needs to be global too. Infrastructure that only works where internet and power are reliable, or that's designed for one particular country, contradicts open licenses.
Delmar Larsen's call for a National Open Courseware Ecosystem gets one thing right: open education needs to become an OER ecosystem, not another collection of repositories. He correctly points out that the systems that make open content usable at scale, including homework and assessment, gradebook integration, authoring, analytics, interoperability, accessibility, AI, and sustained technical and teacher support, also need to be open.
But his national, higher-ed frame misses an opportunity. K-12 is the real test, and it's global. A system centered on North American higher education starts with a population that's already relatively well connected and surrounded by digital infrastructure. That's not where the hardest design problems are, and it's not where we should start building a new OER ecosystem. K-12 is where foundational digital skills develop, where teacher capacity matters most, and where access decisions shape a learner's entire trajectory. It's also where the world's largest population of learners is found, from underfunded American districts to classrooms in Tamale or Manila. Solving the hard problems is what generates funding.
The design test
The design test can't be: will this work in a well-connected university? It has to be: can this work in an elementary classroom with no internet? If it works without internet, the system scales upward into better-resourced environments. If it doesn't, it isn't really an open system, and it isn't very usable for the largest group of learners.
There's a practical reason to frame the project nationally: national priorities attract national money. The NSF, the Department of Education, state systems and their legislators come with identifiable institutions, constituencies, and budgets. "Global open courseware infrastructure" doesn't come with an obvious state or federal line item.
That makes a national strategy understandable; it doesn't make it the right place to start. If the funding source quietly determines the user population, the ecosystem gets optimized for the institutions that finance and govern it, while everyone else remains an afterthought. Open education can't confuse the geography of its funding with the geography of its users.
Delmar's proposed ecosystem assumes cloud-hosted platforms, real-time analytics, streaming content, and continuous integration among services. But much of the world, and increasingly much of U.S. K-12, doesn't operate under those conditions. Electricity may be intermittent. Broadband may be expensive, unreliable, or absent. Devices may be shared among students or across a whole classroom. Governors are talking about getting rid of anything that uses a screen, except when it comes to testing.
Offline-first can't be a feature bolted on afterward, because it changes the system's fundamental architecture. Connected-first systems prioritize cloud services and always-on integrations; offline-first systems require local storage, asynchronous workflows, low-bandwidth content, and the ability to operate independently of the cloud. The two sets of priorities aren't compatible, though offline-first doesn't rule out connectivity where it exists.
The historical pattern in ed-tech is to build for the well-resourced environment, then treat everyone else as a special case. A better approach designs for the harder problem first. An offline-first system still works beautifully when broadband and electricity are abundant; a cloud-dependent system can't reliably move the other direction. Designing for scarcity isn't designing for less: it's designing for resilience.
Capacity-building is infrastructure
The same principle applies to people. Teacher training, digital literacy, implementation support, and ongoing capacity development aren't things that follow the technology: they're part of the infrastructure itself. A sophisticated platform only people with existing technical skills can use doesn't democratize access; it just compounds the hardware barrier with an expertise barrier.
Capacity-building has to be designed, funded, and staffed alongside the technical architecture, and it has to be global from the start. A U.S.-centric training pipeline doesn't represent the teachers who'll use the system elsewhere. The people building and supporting it need to learn from classrooms facing the greatest constraints, not from the institutions with the largest budgets. The neighborhoods around U.S. universities are not good testing environments for ed-tech.
ADAPT shows why this matters
LibreTexts' ADAPT integrates WebWork, IMathAS, H5P, and native QTI into a single homework platform, but it doesn't work in a classroom with no internet. QTI and H5P can port into Moodle relatively cleanly, but WebWork and IMathAS depend on server-side infrastructure for randomization and grading, and features like Learning Trees, analytics, and clicker integration have no straightforward offline equivalent. This isn't a criticism of ADAPT's engineering; it illustrates the consequences of its design assumptions. Moving a platform built for connected higher ed into an offline K-12 classroom isn't a matter of changing settings, it requires architectural change. Without a radically different approach, a new ecosystem is very likely to stay stuck in the current model.
Open content hasn't displaced the infrastructure underneath it
In the current model, an open textbook can replace a commercial one, but the learning still depends on Pearson's, Cengage's, or McGraw Hill's proprietary assessment and grading systems. The textbook is free; the infrastructure that enables the learning isn't. Commercial platforms still largely control assessment, grading, analytics, and increasingly the data students generate. In K-12, Google controls most things; Google is "kinda free," but definitely not open.
The goal isn't more open content; it's open learning infrastructure, where content, assessment, data, and delivery can operate without a proprietary layer underneath, whether the classroom has gigabit broadband or none. Otherwise open education just reproduces the dependency it set out to remove: the vendors change, but the dependency remains.
This is where licensing choice stops being a side issue. A CC BY license permits commercial reuse, which sounds generous but leaves the door open for a for-profit company to repackage openly licensed content behind a proprietary platform, so the ecosystem's usability once again depends on a vendor. CC BY NC closes that door. It keeps the content free to adapt, remix, and redistribute, while blocking a commercial actor from enclosing it inside a paywalled homework system or subscription analytics dashboard.
A non-commercial license means the infrastructure built around the content (authoring tools, assessment engines, offline sync layers) stays available to the public institutions, NGOs, and ministries of education that actually need to deploy it in low-resource classrooms, rather than becoming a licensing fee for a startup that spots an underserved market. It also protects the incentive to build shared, global infrastructure at all: if any contributor's improvements can be commercially captured and resold, the case for pooling effort into one open system weakens. CC BY-NC keeps the value inside the commons.
I'm not arguing against sustainability funding, paid support contracts, or vendors building compatible tools on top of an open core for those who can afford it. I'm arguing that the core itself, the content and the infrastructure layered directly on it, has to stay open, free, and safe from being quietly re-enclosed the moment it becomes valuable enough to enclose.
The opportunity is bigger than open courseware
The opportunity isn't a better repository of open courses for U.S. higher ed. It's a genuinely open learning infrastructure that can serve a classroom anywhere: global by design; offline-first, not offline eventually; K-20-centered, not higher-ed-centered; open beyond content, including assessment and analytics; capacity-building treated as infrastructure, not an afterthought; and locally adaptable in language and curriculum.
National funding, state systems, universities, and foundations can all contribute as they're able and willing, but none of those funding sources should define who the system is for. The assumptions built into the architecture on day one determine who can easily use it. A national system that might eventually go global will carry assumptions that require expensive retrofitting later; a global, offline-first, K-20-centered system can be adopted nationally just as easily, and has far larger reach, and is more likely to attract the necessary funding.
OER already solved the problem of crossing borders; its licenses do that automatically. Now, the infrastructure needs to catch up.
No comments:
Post a Comment