> For a crowd sourced dataset, a strict ontology anyway wouldn't work. Instead of messy tag definitions you'd have tag use that didn't align with the definitions.
I don't really agree. Wikipedia has clearly defined categories (with a very deep hierarchy), and they make it work. There is constant effort in recategorising, but there are tools to support it. No one is saying "oh, just tag stuff however the hell you feel like it".
> The tagging system in the iD editor tries to address the issue, supporting search terms and suggesting related tags and so on.
It also introduces problems in that it takes a particular interpretation of each tag which isn't necessarily right everywhere. Like calling highway=track an "unmaintained track road", where this concept of "unmaintained" doesn't come from the wiki anywhere - they just seem to have invented it.
>The article is more about the underlying storage of the geometries (I don't think there is the same level of interest in changing the basic approach to tagging/categorization).
There are big issues in this too. Particularly the lack of distinction between line and polygon features, which the consumer is supposed to infer from the tags (building=yes is a polygon, highway=pedestrian is a linestring, a circular walking path), with area=yes used in the most ambiguous cases. Plus all the mess with relations, super-relations etc.
