Agreed, I don’t know why Mateusz tried to steer this conversation towards “we should remove names in favour of brands”, that was not the goal of the discussion. Let’s focus on the proposal.
…which will never end. There is no fixed truth here, and consensus… well, opinions are as divided as I am, internally.
If I were a data consumer, I would use a logic like: if exist: name, else brand space branch. Can’t miss, right!
1 Like
OK, but what is the current proposed text referring to brands? The linked page still says without qualification that brand names should not be used, which contradicts actual practice in many cases.
(Perhaps it would also be useful to clarify if you are trying to document how brand names are used in the name tag, or to propose a change).
1 Like
Can we start from the common understanding that brand names are not proper names? I am assuming this now. Otherwise, please address it here.
Historically, the tag name is probably one of the oldest tags, it was already in use more than 10 million times in 2008. Keys such as operator, owner or brand only became widespread later (OSM Tag History). The brand key, for example, only really became established in 2018.
For these early years, the recording of brand names with the key name is therefore justified and easy to follow.
Today, it goes without saying that the brand name is recorded with the brand key. However, there is a ‘convention’ of also adding the brand name in the name tag if the feature does not have a proper name or the proper name is unknown (Distinguishing between brand, name, operator, etc.) with the following justification:
“As the name tag … is supported by almost any application”
Isn’t this a case of ‘tagging for the renderer’?
Some problems come from the default settings of the iD editor. This charging station, for example, with brand=Tesla Supercharger
gets default name=Tesla Supercharger and short_name=Tesla.
I think that’s nonsense.
Most of the name=brand entries probably come from NSI. I’m not sure how many mappers still do this out of conviction.
Strong words that don’t correspond to my understanding of the Wiki at all. In my opinion, the wiki does not say right/wrong, but recommends. Of course you can deviate from this, but it should be well justified.
My consequence for the key brand would be an addition:
Some mappers also add the brand name in the name if there is no proper name. This historical and/or local convention should be observed but also questioned, as it contradicts the definition of the name key.
I will also insert a corresponding note in key:name.
3 Likes
I would put it slightly differently: without the separate discussion (which is a good idea) the documentation for the name key should not say whether or not it should be used for brands.
4 Likes
I genuinely don’t see why a brand name can’t be the proper name of a point of interest.
If a Burger King restaurant presents itself as simply Burger King, and that’s how people refer to it in conversation, for me its name is Burger King. That’s how people tell it apart from other businesses on the street including the KFC next door, consistent with the proposal: “Proper names name a single object in a group of similar objects.”
I agree that if a restaurant is publicized as Burger King X, where X is some kind of geographical or other identifier, that may be a more appropriate name. But around me, they don’t do that. If a distinctions needs to be made between two Burger Kings it would be done with the street address or the name of the enclosing shopping centre if there is one.
Having said all that, none of this really seems relevant to simply restructuring the page. As suggested by @Wynndale, it could be avoided (or deferred to a future discussion) by simply not saying anything about using brands as names.
11 Likes
I did question this assumption earlier but can elaborate. Proper name continues to be a problematic term for the notion you want to express, not least because brand names are some of the most obvious examples of proper names – in a grammatical sense. “McDonald’s” is a proper name; “hamburger” is not. This is not a matter of debate in the real world.
Instead, you’ve latched onto the philosophical sense of this term, even though the ideas of John Stuart Mill and Bertrand Russell are so far removed from the day-to-day concerns of us simpletons who map and use maps. Ironically, philosophers of language would be just as quick to claim a “descriptive” or “constructed” name such as “Chicago rat hole” as one of their proper names.
Judging from your counterexamples, you seem to be most concerned about redundancy between name=* and other keys such as brand=*, operator=*, or denomination=*. There are plenty of situations in which name=* should not duplicate these keys, but human language is complex enough that enumerating or explaining these situations would undermine the diet that you wanted to put the article through. You seem unmoved by the generic names preferred by many Protestant houses of worship, such as “Primitive Methodist Chapel” and “Kingdom Hall of Jehovah’s Witnesses”. At what point does advocating for cleaner map data cross over into denying on-the-ground reality?
Anyways, name=* is only one of the many name keys. If a naming criterion has any relevance to loc_name=*, official_name=*, or alt_name=*, then it probably belongs on the main “Names” page instead.
I’m glad we’re finally getting to the heart of the matter. These days, this NSI preset no longer locks the Name field, so you have an opportunity to replace this default name with something more unique while adhering to the “ground truth” principle. Go wild! For some kinds of POIs, it even leaves the Name field blank, since no one ever refers to an ATM operated by People’s Bank as “a People’s Bank”, or a Paris Metro station as “a Paris Metro”.
Should NSI treat Tesla Superchargers similarly? I suppose it depends on whether you view one of these units as merely a piece of mass-manufactured road furniture, akin to a fire hydrant, or whether you view it as a destination unto itself, akin to a gas station. At least from what I’ve seen, living in an area with a high rate of Tesla ownership, the brand recognition here is very strong: people only refer to “a Tesla Supercharger”, never for instance “a Tesla Supercharger charging station”. That said, would they refer to it as just “a Tesla” for short, matching both the short_name=* tag and the wordmark on the front of the unit? Probably not. But this discussion is about name=*, not short_name=*, right?
5 Likes
Maybe? I am not really sure what “proper name” is. But this is an extremely strong argument against redefining name key as proposed in this thread.
And redefining one of fundamental tags would need massive support. Much greater than wiki page shortening, without meaning change, would need.
name key is for commonly name used locally.
And yes, in many cases it is brand name.
Speaking from local experience (that I am pretty sure it’s widely shared): when I cycle to one of nearby supermarket then I go to Lidl or Biedronka. Yes, it is their common name. Yes, everybody else is naming it this way.
Yes, it is their common local name. Yes, it should go into their name tag. Yes, it was always tagged this way and it predates NSI.
What was described as “proper names” for shops in Poland is not even name, using it would be absurd and harmful. And would violate verifiability guidelines. Seriously, it was claimed that address description should go into name tag.
(Yes, brand is not always the same, nearby milk bar* is named “Centrum” or “Bar mleczny Centrum”, operated by Społem coop, Społem branded)
(Yes, NSI and blind upgrading in iD caused some damage - but it has not really changed or established this tagging practice, at least in Poland)
*very cheap (3 euro for full meal with 2 courses) fast food place serving tasty traditional dishes, partially funded by government.
(Not sure about Tesla Superchargers - maybe it should go to description tag, maybe name)
6 Likes
Not using “proper name” also seems needed, as it apparently leads people to making this claims.
1 Like
Historically, the tag name is probably one of the oldest tags, it was already in use more than 10 million times in 2008. Keys such as operator, owner or brand only became widespread later (OSM Tag History). The brand key, for example, only really became established in 2018
in 2008, there mostly it was roads that were mapped, those 10 million names definitely weren’t on shops: http://taghistory.raifer.tech/#***/shop/
brand is also quite old as a tag, saying it only was really used since 2018 doesn’t nail it, although in 2018/19 there is a significant bend in the graph, the tag was already established for much longer, initially for car dealerships and petrol stations: http://taghistory.raifer.tech/#***/brand
5 Likes
Nobody is trying to redefine names, except the ones who try to make it a name=brand only.
namekey is for commonly name used locally.
Not locally necessarily. Usually it’s the local name which is best, but it can be also the name used by people familiar with the feature, or by the most people. Again, that’s not a redefinition.
It is not. As always, you are conflating the fact that people use a brand, by convenience, to refer to something, and the name of the feature.
What was described as “proper names” for shops in Poland is not even name, using it would be absurd and harmful.
It wouldn’t. Use the real name if it exists, and by convenience for downstream support, use name=brand if no specific name is used. It’s what is used almost everywhere since the beginning of OSM up to today.
We are not talking about Poland. We are talking globally. Real/proper names are preffered to brands globally to brands only, but they can still be used. Even in Poland, the examples you used show that some chains do use public and distinct names, such as Ikea.
Yes, NSI and blind upgrading in iD caused some damage
It did a lot of damages, and continues to do so, at least in France.
1 Like
Yes those stations have public names, just like gas stations. The NSI preset should be changed.
That’s a side-effect of the NSI project goal of setting as much tags as possible : putting in incomplete name is preferred to setting the correct name.
1 Like
So if I’m going down to Aldi to shop, where do you want me to say I’m going?
1 Like
You can say anything. I don’t know what the name of your shop is, or whether you prefer to use brands or names to refer to shops.
However I know the names of McDs around me, even if I say that I go to “McDo” for convenience.
I know the names of multiple cinemas, even if I only usually say that I’m going to “Pathé”/“CGR”.
I say that I go to “Auchan”, even though if you ask me what is the name of this shop, it’s “Auchan X”.
Not sure what you are trying to say, apart that you prefer saying the brands of shops instead of their name. It’s totally fine, most do too, including me ! And you can display whatever you want on your OSM applications.
However, in the OSM database, the name of shops should be used as names, if presented to the public. And brands as a fallback if names are not shared, to help OSM usage. That’s what it always has been, in OSM, most of the web and world.
Edit : Tagging both the name and brand allows to best reflect what’s on the ground and digital, while still allowing each OSM user/consumer to make its own choice. Instead of forcing this choice and removing real-world data from OSM.
1 Like
I think it is core disagreement.
I think that we right now use for name a name that is used by people to refer to specific shop. Even if it is the same as brand or name for other locations.
In most cases it is the same as signed name - and for chain shops, typically the same as brand.
I think that using shop title appearing only in catalogue on app or website, not signed, not actually used by people is not fitting for name tag even if it philosophically is “proper name”.
As I understand, you claim that situation is reverse and that current use is effect of damage caused by NSI.
Have you tried looking at map data before NSI existed to confirm this claim is accurate?
Or are you claiming that name tag was always used wrongly in the same way?
1 Like
Might also be worth adding that the original NSI data was based on what already existed in OSM.
1 Like
What are the “names” of the McDonald’s around you?
Even though I might say that I’m going to “Mickey D’s”, “McDick’s”, or “Rotten Ronnie’s” ( 
), the restaurants near me are named simply “McDonald’s”.
4 Likes
The shopping centre near my house has many stand-alone shops, and also branches of shop chains. All of the branch shops have the brand name on the entrance, most have a qualifier with it such as the location or the local owner/operator.
How are they named by the people? As depicted at the windows/displays/entrances; same as for all the other shops. Locals make no difference. The qualifiers are seldom used; except if there is another branch nearby e.g. in the next shopping centre in the same town.
So the brand name equals the commonly used name. If the name tag should hold the commonly used name, then it equals the brand name in most cases. Sometimes the branch can be added, according to the map what it says on the window rule.
So, AFIAC the choice is not whether or not it IS a name; the choice is whether or not the brand and branch could be repeated in the name tag. To me, that is a mapper’s preference. I don’t think it can be settled here.
1 Like
Does this real-world data need to be in name=*, or can it be in something else like alt_name=* or official_name=* that better communicates its obscurity?
One of the downsides of insisting on maximal specificity in name=* is that many things that laypeople consider to be named would go without an explicitly tagged name, and would even have noname=yes on them, contradicting the facts on the ground.
What are the public names of these charging stations, and how would a mapper obtain this information? As far as I know, in a set of all the charging stations in a city, the philosophically correct proper name would be “Tesla Supercharger”. Meanwhile, in a set of all the charging stations within that brand, I have no idea if Tesla somehow communicates the individual stations’ names to their customers, but they don’t do so on site or through their website.
Some gas stations do have well-known names that differ from the brand. For example, this Marathon station is named “Obermeyer Marathon” after its owner, as the sign says. It’s even listed in the National Register of Historic Places under that name. “Blue Ash Shell” is well-known because of the attached car repair shop by the same name. But at least in that region, these are rare exceptions rather than the rule.
Consider this more representative example that I gave earlier:
Are you saying that this ordinary gas station can only be named if someone goes there and pays for gas by credit or debit card? And that the motoring public, not just city inspectors, need to know that it’s called “N.B. Oil Company Incorporated #1” above all else? As someone who habitually types curly apostrophes where a straight one will do, I certainly sympathize with the desire for precision. But OSM will be OK if we relegate this minutiae to official_name=*, as I did in this situation.
2 Likes