Fkv voted no on Blood donation 2 because he believes the
donation:compensation:* keys will cause inconsistencies, e.g.
donation:compensation=no together with
donation:compensation:payment=yes. I can see his point though I don't
think it would happen often.
He proposes to use a semicolon-separated list, e.g.
donation:compensation=payment;vouchers.
We both know this is controversial, but why in fact?[1] It is supposed
to be hard to interpret for data consumers, but can't they just as
easily make a list with booleans of all the values that are present in
the tag?
1: donation:compensation=yes donation:compensation:payment=yes
donation:compensation:vouchers=yes
2: donation:compensation=payment;vouchers
-> "compensation": true, { "payment": true, "vouchers": true }
(possible result in the memory of a parser)
1: donation:compensation=yes donation:compensation:vouchers=yes
2: donation:compensation=vouchers
-> "compensation": true, { "payment": false, "vouchers": true }
1+2: donation:compensation=no
-> "compensation": false, { "payment": false, "vouchers": false }
Is there perhaps a significant performance drawback to having to parse
semicolon lists or what am I missing?
[1] wiki.osm.org/wiki/Semi-colon_value_separator explains you should
use the namespaced approach, but not why
2015-07-12 20:12 GMT+02:00 Ruben Maes <ruben.mae...@gmail.com>:
> Voting is now open on the modified proposal:
> https://wiki.openstreetmap.org/wiki/Proposed_features/Blood_donation_2
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging
