The brief
Navigation apps are very good at getting somebody from A to B without wasting time. Leisure walking inverts the objective. The point is not to arrive efficiently, and the questions a recreational walker actually asks are ones no routing system can answer: is this scenic or bleak, is it safe alone after dark, does it pass anywhere worth stopping.
Those qualities are real, and they exist nowhere in standard geospatial formats. The pipelines every walking app depends on, including road networks, points of interest and terrain models, were built for logistics. Asking them about the character of a place is asking a question the schema has no field for.
The working group set out to give the field the shared reference it was missing, on the argument that the barrier to better walking applications is no longer computational. It is that everybody building one starts from scratch, invents their own vocabulary, and produces something the next team cannot build on.
The insight
The useful move was to stop treating this as a research problem and treat it as a translation problem.
There is no shortage of relevant theory. Environmental psychology has decades of work on how people respond to places. Space syntax formalises how spatial configuration shapes movement. Platial GIS has been arguing for place over coordinates for years. None of it arrives in a form a developer with a sprint deadline can use, so in practice none of it reaches the applications where it would matter.
This was not an abstract observation. WalkGrid had spent four years demonstrating both halves of the problem, first that environmental preference can be modelled and routed on, as set out in the thesis, and second that every piece of it had to be built from scratch because no shared foundation existed.
So the outputs were built backwards from the point of use. Not a paper arguing that place representation matters, but a checklist that tells you what to do in the design phase and what to test before release. Not a call for better evaluation, but an evaluation framework somebody can run. The theory is all still in there, and it is packaged as instructions rather than as argument.
Publishing it openly and outside an institutional repository was part of the same reasoning. A local authority team or a two-person startup is exactly who needs this, and exactly who will not pay for journal access or know to look.
What it does not do
It sets no standard. The representation gap is named, described and argued about, but the group did not produce a schema, and until somebody does the underlying interoperability problem remains exactly where it was.
The documentation is a synthesis of existing work and expert convening rather than new empirical research. That was the right shape for a three-month impact grant and it means the frameworks here are reasoned rather than validated. The Planning Checklist has not been through a controlled trial against teams who did not use it.
Adoption is unmeasured. The platform is public and built to persist, but the project ended before there was any evidence about who found it or whether the applications built afterwards were better for it.
