Open data is one of the most hopeful and most disappointing movements I have been part of. The premise is right: public data is a public asset, and the public has a right to use it. What went wrong was almost entirely execution and sustainability.
Publishing a CSV on a portal is not a service. It has no contract, no versioning, no uptime commitment, no support channel, and frequently no owner once the grant that funded it ends. Developers build something, the file format changes, the URL moves, the portal gets redesigned, and the thing they built quietly dies. That happened enough times that a lot of good people stopped trying.
The gap was always accessibility rather than availability. Opening up information is not enough — it has to be usable, which means APIs, documentation, stable identifiers, quality commitments, and someone whose job it is to keep it working.
The analogy I keep returning to is public lands. We do not merely declare a park open; we fund stewardship, maintain the trails, and staff it. Public data needs the same, and it almost never gets it, because the political win is in the announcement and the cost is in the decade afterwards.
