Talk:Relations/Proposed/Associated Entrance

From OpenStreetMap Wiki
Jump to: navigation, search

This is looking good and I think it's a useful proposal.

--Polyglot 22:49, 23 April 2011 (BST)


seems like we had the same kind of idea at the same day. though i think not quit the same, since yours covers physical entrances, mine covers addresses: so, i can see situations, where both kind of relations are needed at the same time (for example when there are multiple POIs for one address that isn't assigned to a specific entrance) --Flaimo 11:47, 24 April 2011 (BST)

The idea for a relation of type associatedEntrance is a little bit older and has already been discussed here (in German). The reason for inventing it was to avoid redundancy in the assignment of addresses to POI -- exactly the situation you describe in your proposal. And the solutions we describe in our proposals seem to be quite similar as well: The nodes carrying the housenumber in the pictures of your proposal are entrance nodes. Differences are that my entrance-node members have a role and that I allow for more than one entrance in the relation of type associatedEntrance. This seems to be closer to reality, since a shop in a larger building may indeed have more than one entrance. In your proposal you say that an address need not be assigned to an entrance-node. Whatever node this might be, it would not be in compliance with the Karlsruher scheme, as far as I understand it. Do you know of any proposal describing such a node? If not, the next step should be to create a proposal for such a node -- or better -- to enhance the Karlsruher scheme accordingly. Otherwise the notion of such a node would be too vague and I would not count it. -- Schlauchboot 08:31, 25 April 2011 (CEST)

associatedEntrance vs associated_entrance

Why camel case? Separating words with underscores is a much more common convention for OSM tags. -- Binary Alchemy 05:39, 25 April 2011 (BST)

I had the Karlsruher scheme in mind, which uses associatedStreet. But this does not really matter. One could use associated_entrance instead of associatedEntrance. Since this seems to be true for relations, I changed it. -- Schlauchboot 08:36, 25 April 2011 (CEST)
i would rather stick with camel case so that all associated* types are written the same way. --Flaimo 11:57, 25 April 2011 (BST)
I prefer underscores. There is exacly 1 popular tag using camel case: type=associatedStreet. That's hardly a reason to introduce more exceptions to the general rule to use underscores. If we want consistency, we should rather change associatedStreet to associated_street, too... --Tordanik 14:19, 26 April 2011 (BST)