PAA schrieb: > Using the proposed shop=pet as an example: > * grooming=yes: The shop offers pet grooming services > * kennel=yes: The shop offers kennel (pet keeping or tending) services > * training=yes: The shop offers obedience training services > > Or > > * service=(grooming;kennel;training)
I suggest using the first variant. It's easier for existing tools to evaluate, because they usually can't handle semicolon-separated values (these are just a relatively common convention, after all). Also, duplicates are prevented automatically. Additionally, separate keys are better if you want to use implied values. That doesn't make much sense for your example, but it's an advantage of that variant in general. If pet shops e.g. usually offered a certain set of services and that was a widely accepted fact, it would be possible to make that information implicit. With the second variant, it then wouldn't be possible to remove one implied service for a certain shop (due to the lack of a =no option). No matter which variant you choose, you might want to add a prefix/"namespace" to those keys, which was done e.g. for recycling keys: http://wiki.openstreetmap.org/wiki/Tag:amenity%3Drecycling It helps to avoid key name conflicts. (I'd expect "training" or "service" to be used in other contexts, too, which would make validation and documentation tricky.) It also is convenient when keys are sorted alphabetically. I'd probably use pet:* rather than of shop:pet:* because the keys would be useful for other providers of pet related services, too, not just shops. Tobias Knerr _______________________________________________ talk mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk
