Transparency Is a Right.

A freedom-of-information law passes. A portal launches. Someone counts the records now "open" and writes a press release. That's disclosure: an event, staged once, over the moment the cameras leave.

Transparency is a right, not an event. It's a standing claim: the ability to search a record, verify it, cross-reference it, catch the lie in it. Anything short of that isn't transparency, no matter how legal it is.

That gap between disclosure and transparency shows up in three concrete places.

Start with the software itself. A city publishes its data through a proprietary platform, so the tool actually producing the public record stays closed even though the output looks open. Open-source alternatives exist specifically to close that gap, but most agencies never switch, and when the vendor contract ends, the data can disappear along with it.

Even when the platform isn't the problem, the portal often is. A study of 280 government open-data portals found that 56% were "pretender" portals: technically online, legally compliant, and useless for anyone trying to export the data in bulk or reach it through an API.

280 portals, one study

56% of government open-data portals studied were "pretender" portals: technically online, but you can't export the data in bulk or reach it through an API.

And underneath both of those sits a budget problem. The EU spends an estimated €264 billion a year on public-sector IT, most of it on proprietary products and services that deepen vendor lock-in rather than fund something open and inspectable.

EU public-sector IT spending, per year

€264B

Spent every year on EU public-sector IT, predominantly on proprietary products and services, per the Free Software Foundation Europe's estimate.

Mir Labs works on these three problems directly, but not because we think we're the only ones who should. There are other people doing this work, sometimes at another agency, sometimes as an independent developer or a civic-tech group with none of the budget an institution has, running the same open-source stack a big agency runs, CKAN, DKAN, the libraries underneath them, with far less capacity to keep it patched and audited. Part of the point is making their job easier too: building tools they can adopt instead of rebuilding, and publishing audits and datasets they can cite instead of re-running the same analysis from scratch.

So we build open-source, self-hostable tools that replace the closed layer, so the software producing a public record is as inspectable as the record itself. We audit whether portals and pipelines actually work, not whether they technically launched, and say plainly which ones are pretenders, and we audit the open-source packages those portals and pipelines are quietly built on, because a portal is only as trustworthy as its dependency tree. And we ship datasets and tools that are already built, cleaned, versioned, and free, so nobody, us included, has to fund a proprietary system just to get the same result next year.

The pattern isn't unique to government. Research data "available upon request." Corporate filings in formats built to resist parsing. Standards behind a paywall. Open in principle, closed in practice, for the same reason every time: someone already had their disclosure event and stopped paying attention.

If there's a public record you can see but can't use, that's the queue. It's open.