I found the added complexity of this normalization funny/ironic when I wanted to process some OSM data. As IMO, the benefits don't nearly outweigh the added com

I found the added complexity of this normalization funny/ironic when I wanted to process some OSM data. As IMO, the benefits don't nearly outweigh the added complexity and costs caused by it. A different way of creating new topological connections wouldn't be that difficult. For example, some topological connections are already being created via relation type. I'm not involved with OSM project, but I think that some changes to OSM data model would benefit it greatly. Resolved geometries would enable stream processing and decrease complexity for majority of task. It would make OSM data much more accessible.