Proposal talk:Payment denominations
Duplicate / overlapping
Proposed features/coins / notes acceptance with cash_*:*=* is explicitly provided for amenity=vending_machine as well. Mainly generalizing cash_in=* from amenity=atm. --- Kovposch (talk) 20:26, 24 September 2022 (UTC)
- I know, I have mentioned it under "rationale" and tried to explain why my proposal is different. Discostu36 (talk) 20:29, 24 September 2022 (UTC)
- But how do you handle
cash_in=*inamenity=atm? It's not "payment". Will end up with 2 parallel solutions. Another semantics issue is convenience stores accepting the same cash for topping up stored-value cards, and paying bills. "payment" doesn't fully describe the situation. ATM andamenity=payment_terminalallows you to pay bills too --- Would ultimately I add bothpayment:*:coins=*andcash_in:*=*(for the latter, eg topping up pre-paid SIM) or equivalent to cater both paying and depositing? Not to mention post offices providing banking service.
What's the difference withcash_in:*=*then? It's as established as adding the new suffixpayment:*:denomination=*anyway. If you want compatibility, you could instead add them directly inpayment:coins=*andpayment:notes=*. For you comparison ofpayment:coins=5c;10c;20c;50c;1€;2€, one could equally add the currency used in otherpayment:*=*without a denomination, especially inpayment:credit_cards=*. This will be more specific thancurrency:*=*. So cash is not unique in extensibility.
Indeed your use ofpayment:coins:*=*andpayment:notes:*=*does conflict with the format in Key:payment#Currencies. The possible combinations are less numerous than thepayment:notes:10=*you compared. Which is something I commented on in the other proposal: Repeating the currency code in every denomination seems annoying. Using*:coins:EUR=0.05;0.10;0.20;0.50;1;2saves some typing effort and the risk of typos. One advantage the other proposal has is the provision of omitting the currency code, with*:coins=0.05;0.10;0.20;0.50;1;2defaulting to the local currency.
---- Kovposch (talk) 21:09, 24 September 2022 (UTC)- Thank you for your feedback. I have to admit that I don't understand all the points you are trying to make but let me answer to two of them:
- 1. This proposal is not trying to solve overlap between
payment=*: andcash_in=*. It is an extension to the payment: scheme and can be used everywhere where payment: is used. At the moment, cash_in is almost exclusively used on ATMs, so the overlap is also not very big. - 2. Your point of this proposal being add odds with how currencies are currently handled in the payment scheme is correct. I will have a look at that. Discostu36 (talk) 08:44, 25 September 2022 (UTC)
- I have decided against using the sub-key solution as suggested in
payment=*as it is not commonly used, makes the key even longer / more complicated and using the data harder. Discostu36 (talk) 20:35, 25 September 2022 (UTC)
- I have decided against using the sub-key solution as suggested in
- But how do you handle
"This information should not be in OSM"
Continuing the discussion in the voting started by @Freetz::
As can be seen in this (incomplete) table, payment denominations are already tagged by users with different mapping schemes. This proposal is not about "should this information be in OSM" because it already is. As OSM doesn't know a distinction between important and unimportant information, there is also nothing that could prevent users from entering it. The only thing that this proposal will change is that (hopefully) all users will use the same tagging for that information in the future, making it easier for data consumers (if interested) to use the gathered information. Discostu36 (talk) 20:32, 16 October 2022 (UTC)
- I just checked
coin:type=*on overpass. Out of the about 20 POIs I randomly clicked on (in about 600 total), 100% areamenity=vending_machine+vending=elongated_coin. Maybe there are different things tagged, but statistics suggest that by far the most use is withvending=elongated_coin, most likely becausecoin:type=*is listed on its wiki page. The 110payment:coins=5c;10c;20c;50c;1€;2€mostly occur around a single town in Germany. I haven't checked, but I wouldn't be surprised if only single mapper did that. Pretty much the same seems to be true for the 100payment:notes:max=*occurrences. - I don't think it can be concluded that denomination tagging is currently done by many mappers and needs to be formalized. It's rather done by a few individuals only. I suspect that it's highly unlikely that there is any data consumer application looking for denomination acceptance information.
- While of course anybody can enter pretty much any data in OSM, a formal approval of a scheme for a specific kind of information declares this information is relevant (at least appriciated) and probably encourages mappers to enter such data. I think
vending=elongated_coinis a good example for that. I believe it is better to not promote denomination tagging and to not encourage mappers to tag it with approval of this proposal, or any other mapping scheme. Denomination tagging will end up in poor data quality after a while, and regardless of the importance of denomination data, it will be not dependable, hence useless. - I have recently noticed many outdated retail amenities (type and opening hours) in downtowns of several large cities in different countries. Given this experience, I wish local mappers take more care about maintaining the correctness of (importand data in) existing amenities rather than including additional data nobody will ever maintain. Let's concentrate on data quality, not on amount.
- --Freetz (talk) 02:39, 17 October 2022 (UTC)
- Btw, it should be pointed out that
coin:type=*inamenity=vending_machine+vending=elongated_coinis not for the payment coins but for the coin that is crushed. So it is used for the description of the product sold, not of the payment. - --Freetz (talk) 03:11, 17 October 2022 (UTC)
- Btw, it should be pointed out that
- All 600 elements with
coin:type=*are in factamenity=vending_machine+vending=elongated_coin. As this is not a payment tag, this tag and its use are not relevant for this proposal. In conclusion, round about 200 elements appear to be tagged with payment denominations, most of them by a few mappers. - --Freetz (talk) 03:25, 17 October 2022 (UTC)
- All 600 elements with