Multiple delimited names in the name tag

Minh_Nguyen · 2023-01-24T03:09:06.693Z

I’m glad to see some experimentation on the osm-carto front. I look forward to some of these experiments being made available for mappers to experience firsthand.

What @imagico has prototyped needn’t be mutually exclusive to what Americana has implemented. It’s entirely possible that a perfect solution in a renderer would require the combination of multiple approaches, but realistically some data consumers will need something in between that and the most basic handling of name.

Hungerburg:

From reading there, mappers already do as I proposed, by turning the default_language key into such a list, of course separated by semicolon.

As always, the devil is in the details. So far this approach relies on some details that hopefully can be clarified by the time it goes into production. Here’s a few things that come to mind:

default_language itself was rejected in 2018. There seems to be some enthusiastic support for it in this thread, so I wonder if that vote’s participants were missing important context. It may be an uphill battle to convince editors to support a rejected key.
It would be essential to mitigate the risk of vandalism or accidental breakage. Boundary relations break all the time in a manner that is often difficult to fix without local knowledge. To mitigate this risk, the FAQ points out that boundary-based defaults could be subject to a delay, similar to the changeset review process implemented by some data consumers such as Mapbox. I’m intrigued by this possibility but also apprehensive about introducing another process similar to the coastline process that works except when it doesn’t.
Since language use does not neatly correspond to administrative boundaries in the real world, this prototype relies on arbitrary features to be tagged with default_language. This is equivalent to another rejected 2017 proposal in all but name.
Since a space is typically used as a delimiter in North Africa and Hong Kong, the prototype tries to detect boundaries between different scripts. This is naturally only a rudimentary implementation with room for improvement, but I suspect it would be the most fragile aspect of this approach.

The Unicode standard comes with an algorithm for script boundary detection as part of text segmentation; however, there are plenty of edge cases in POI and place names that this algorithm would consider to be degenerate. One challenge is that many common characters are script-neutral, such as “15” or “131”. But clearly Latin letters like “E” can also be part of a non-Latin name. The Japanese community has even standardized on a mix of rōmaji, kana, and kanji in POI names, based on the literal contents of signs.

To be clear, I think it’s OK for a data consumer to develop sophisticated heuristics along these lines. MapLibre GL/Mapbox GL uses similar heuristics to decide whether it’s appropriate to rotate CJK text vertically. But if splitting strings on semicolons is merely a quick fix for “ugly breakage”, to be viewed askance, then what are we to make of text segmentation heuristics that may not even be reliable?

Hungerburg:

Remains, how to handle the non-language based multiples.

Christoph’s prototype does attempt to address this case too, by requiring the component names to be repeated in alt_name or similar. Unfortunately, none of the *_name keys would apply to the monolingual cases I brought up above, such as the shared dentist’s office in Cupertino, so the prototype would simply hide the label, communicating to the mapper that the POI is mistagged. I’m not optimistic that any proposal to redefine alt_name would ever fly, but there has been a suggestion to introduce a new key for this purpose:

Multiple delimited names in the name tag Tagging general discussion

The idea (AIUI) is that it would either be tagged as: display_name=Derry / Londonderry name=Derry;Londonderry Or an alternative proposal: name=Derry / Londonderry names=Derry;Londonderry Amusingly the local vogue appears to be for a tilde, as in Derry~Londonderry. My regex is getting more arcane by the hour.

Minh_Nguyen · 2023-01-24T03:31:51.293Z

In fairness to Christoph, his prototype is primarily aimed at solving the Unihan/Arabic font selection problem, which is a legitimate goal, and only incidentally about laying out compound monolingual or multilingual labels. I quibble with the notion that one goal is inherently more worthy than the other, and especially with a cold calculation based on population, but every software project is entitled to its own priorities. Certainly it’s good to finally choose a Chinese font for Chinese and a Nastaliq font for Urdu, but I don’t see it as inextricably tied to other labeling issues.

For Americana, selecting a font for the label has been slightly less problematic than for a statically rendered style. A combination of happy accidents results in the style automatically selecting not only the user’s preferred language but also, for CJK only, their preferred font in their preferred language. Consequently, the delimiter issue has been more of a focus, but there’s lots of room for improvement around fonts too.

dieterdreist · 2023-01-24T08:49:19.000Z

Firstly, default_language=* may be a fine idea for places where administrative areas have a single primary language, but this falls down completely in numerous places where local language dominance doesn’t follow administrative boundaries. A prime example of this is neighborhoods dominated by ethnic minorities in many of the world’s cities. These neighborhoods have may have one or more local languages that appear on signage, place names, and other features, but which still remain a minority in the smallest administrative area that surrounds them. These neighborhoods also do not necessarily have sharp borders that one could draw a polygon around. Minority-owned businesses named in minority languages may intermingle with majority-culture businesses as well as those of yet other minorities. For example, a predominantly Chinese neighborhood in New York City may overlap with Italian neighborhood, a Korean neighborhood, and a Jewish one and on the street between them one may see an intermingling of businesses with their primary names in 5 different languages and as many or more scripts. default_language completely fails to accommodate this type of real-world complexity

the default language could be tagged on a per object basis. It could either be the standard or be used to override a value from an enclosing polygon.

pnorman · 2023-01-27T22:40:25.513Z

It’s worth remembering that whatever is chosen will need to be acceptable to data consumers who are only looking at the name tag, and no other fields.

Pedja · 2023-01-31T03:22:13.526Z

name tag should be made deprecated with introduction of name:* tags. It just calls for argues.

name:* tags already allow having names in needed languages. Existence of name tag just creates confusion.

Introduction of another tag, something like “local_language” that would contain designation of local language (or languages separated by semicolons) may provide needed information how name is displayed on local signs.

From software side that would clean, straightforward and simple way to tag names in various languages and on social side it would remove arguing and fighting about what should stay in name tag.

name tag lost its basic function to display what is on the ground as that rule is not applied everywhere and is proved to be avoided due to various political biases.

Even where it is agreed how to put multiple languages in name tag it is impractical. Using multilingual name just makes map cluttered. Map renderer has to know local rules and parse content based on those rules.

Deprecating name tag, using name:* tags and using local_language tag would allow universal way to tag names and render maps, by having better control of rendering, regardless of languages used for naming.

stevea · 2023-01-31T03:38:10.998Z

I don’t want to sound discouraging, but deprecating name=* seems like a total non-starter. It might be perhaps the oldest, most consistently used (paired) tag in OSM. It is likely supported by every single use case (router, renderer, text-to-speech module…) out there, and all would need to be re-written / modified if we were to deprecate name=*. I doubt this would happen without great upset to our project.

Modifying the name=* tag, like with name: subtags, well, we’re listening and discussing, as that could actually fly.

This is simply one person’s opinion (mine).

Pedja · 2023-01-31T03:52:43.926Z

It does not have to be removed. It may be generated, based on other rules that are described in a way that allows generating content for name tag.

stevea · 2023-01-31T03:53:25.515Z

Hm, that is a helpful clarification, and I didn’t think of your answer quite that way. Thank you!

Minh_Nguyen · 2023-01-31T08:49:23.828Z

Welcome to this discussion! As you can see, this discussion has dragged on quite a while and is a bit difficult to follow. Your point about the importance of name:* is well taken, but various posts above have touched on why it’s insufficient to rely on a key that indicates which name:* to display. Here are some links to individual comments as a starting point, to avoid repeating these arguments:

The local names are useful even when the map primarily labels the user’s preferred language. OSM Americana demonstrates that the clutter is manageable as long as names can be deduplicated reliably.
We cannot completely avoid the need to separate multiple names by a delimiter. Names in different languages are just a special case of the more general issue that a feature can have multiple names in the local language or multiple names in the user’s preferred language.
A proposal for name:language was rejected in part because mappers didn’t want to repeat the same information on every named feature in a language region.
An alternative proposal for default_language was also rejected. Mappers used it anyways, but its values are often suspect, and it would be an easy target for vandalism or accidental breakage.

If you’re following me so far, what remains is a debate about whether to insert human-readable punctuation or a machine-readable delimiter between multiple tag values when the values happen to be names (but aren’t alternative names, official names, short names, or the names of destinations, brands, operators, or owners).

dieterdreist · 2023-01-31T08:48:46.000Z

From software side that would clean, straightforward and simple way to tag names in various languages and on social side it would remove arguing and fighting about what should stay in name tag.

sure, removing the name tag would stop arguing what to put in the tag, but introducing local_language as a new tag will start new arguing about the exact same topics (which languages to add and in which order)

stevea · 2023-01-31T08:56:24.112Z

Minh, your summary is outstanding; thank you!

Pedja · 2023-02-07T15:33:08.167Z

Sure but it would be lesser issue, and easier to agree or change.

Thing is mappers should be instructed to use name:lang tags more as that would make usage much easier.

okainov · 2024-11-13T14:29:54.319Z

Looked at the recent proposal, I got the problem statement… But I didn’t get how is the proposal going to improve the situation with name alone.
If the feature has all kinds of name:lang tags and has no “name” - all good. But when it has name? Do we expect it to exist in general? Contain multiple languages? This is not clear from the proposal text…

Minh_Nguyen · 2024-11-13T15:43:42.869Z

Which proposal are you referring to? From what I understand, this thread is a debate about how to delimit multiple values in the name=* key. It presupposes that there is a name=* tag, and that multiple languages could be represented in it. There are some other threads about whether to leave name=* unset on certain kinds of features.

Atlantic Ocean: repeated name removal General talk

Over the past several years certain mappers have repeatedly removed the name tag from the Atlantic Ocean node (History Viewer, Deep History). The same pattern can be seen for the North Atlantic and South Atlantic. Proponents of name removal argue that these oceans have no local name because nobody lives there. Regardless of whether this argument has merit, this has a negative downstream effect on OSM based maps. For example, the comprehensive name:lang tags on ocean nodes allows the OSM Amer…

RFC: Intentionally omitted name tags General talk

Comment is requested on a proposal to introduce two tags to indicate the reason why a name=* tag has been omitted from a feature: https://wiki.openstreetmap.org/wiki/Proposal:Omitted_name_tag

okainov · 2024-11-13T15:56:57.050Z

Minh_Nguyen:

Which proposal are you referring to?

The one which mentioned this topic in the email from today Proposal:Add languages: tags for name rendering - OpenStreetMap Wiki

Jarek · 2024-11-13T16:21:38.867Z

okainov:

The one which mentioned this topic in the email from today Proposal:Add languages: tags for name rendering - OpenStreetMap Wiki

To finish crossing the streams, I’m guessing that “the email from today” refers to tagging mailing list message [Tagging] [RFC] Feature Proposal - Add languages: tags for name rendering, and that in turn was prompted by the recent discussion in the Canadian forum Multi-language/default language naming conventions

Minh_Nguyen · 2024-11-13T16:34:26.573Z

There’s a separate thread to collect feedback on the new proposal. Of note for the participants here, the proposal currently doesn’t address the approach of using a semicolon to delimit multiple names in name=*, which was discussed at length here.

[RFC] Feature Proposal - Add languages: tags for name rendering Tagging general discussion

There are a few ongoing discussions on the OSM Community forum about the use of the name tag for features in multi-lingual contexts. The problems stem from treating the name tag as the to-be-rendered localized name. Improvements that address the main concerns could be made by giving renderers enough information in the OSM dataset as to how to select appropriate name: tags for rendering in a given region. Two new tags, languages:official and languages:preferred for use on administrative boundar…

aseigo · 2024-11-13T16:42:39.275Z

Minh_Nguyen:

the proposal currently doesn’t address the approach of using a semicolon to delimit multiple names in name=*, which was discussed at length here.

The reason for that being that multiple names in a single field with a separator (e.g. semicolon):

does not help e.g. audio rendering which benefits from using the right language pack to voice the names. If you’ve ever used a nav set to e.g. English to go through a city in a different language … the results are often less than helpful!
requires changes to be made to all name tags for it to be consistent. Other than being a fair amount of work, even if mostly automated, it makes situations where the policies change very hard to track (see the Haida Gwaii example in the proposal)
it also does not address other multi-lingual use cases, such where the name should appear in one language if its name:<code> tag exists, but then fall back to the official languages of the area, in correct order, if it doesn’t. Yes, this can be done manually by careful editing of all the name tags, but that is very error prone and can not be automated in those cases. (Again, see the Haida Gwaii example in the proposal)
it is not truly backwards compatible. The default case would require changes to all renderers to render the results in a pretty fashion.

In any case, comments and feedback on the proposal from people here who quite obviously care about this issue is very welcome and desired. Hope to see you over there!

okainov · 2024-11-13T16:50:58.889Z

Minh_Nguyen:

There’s a separate thread to collect feedback on the new proposal.

This wasn’t present in proposal wiki unfortunately. But yeah, better let’s move there.

Minh_Nguyen · 2024-11-13T22:47:44.236Z

aseigo:

does not help e.g. audio rendering which benefits from using the right language pack to voice the names. If you’ve ever used a nav set to e.g. English to go through a city in a different language … the results are often less than helpful!

Where I live, a majority of residents speak English as a second language if at all, but all the streets are named exclusively in English. When a Spanish speaker gives directions to turn onto North 4th Street, they may call it calle cuatro in translation or nort fort estret in Spanglish. A Vietnamese speaker won’t flinch at Google Maps calling it nót phót xtrít in Vietrish or Apple Maps calling it đường số bốn in translation.

A decade ago, it was still somewhat rare for a text-to-speech engine to be able to handle embedded foreign-language text, especially without any indication of the language. Today, that capability is quite widespread, even among existing OSM data consumers. For example, the Mapbox Navigation SDK uses Amazon Polly, which has no problem embedding Spanish inside English. If you prefer code switching, then you can mark up the text as a particular language using SSML. In this English recording, the street name “Calle 8” is marked up as Spanish, so it comes out as calle ocho instead of callee eight. This is no longer strictly necessary for common language pairs like English–Spanish, but more distant or obscure language pairs such as English–Chinese or English–Haida could benefit.

I understand that your proposal is designed to make it easier to come up with the language code to mark up the street name. You’re right that a semicolon delimiter doesn’t solve this problem generally. However, your proposal still needs to account for the possibility of multiple names in the same tag, because that can happen even within a single language:

Atlantic Ocean: repeated name removal

Similarly, this reservoir straddling the North Carolina–Virginia state line is also managed by the federal government. It was originally named Bugg’s Island Lake (or Buggs Island Lake due to the federal rule against apostrophes), but Congress later renamed it the John H. Kerr Reservoir after the Congressman from North Carolina. North Carolinians call it Kerr Lake, but Virginians prefer its original name, having tried unsuccessfully to revert the federal name.

Atlantic Ocean: repeated name removal

I don’t know of a waterway with this, but in Ireland, we use regional language suffixes to tag that a certain place has a different name in (Hiberno) English name:en_IE and (British) English name:en_GB. Perhaps that approach could be useful here

Based on @amapanda_ᚐᚋᚐᚅᚇᚐ’s example, I’ve added name:en-US-NC and name:en-US-VA to the Kerr Lake relation (hyphens, not underscores, per the BCP 47 standard). It feels weird to use ISO 3166-2 codes here, as though there are two dialects called “North Carolina English” and “Virginia English”, but it gets the point across. I also changed the name to Kerr Lake;Buggs Island Lake, accounting for the fact that it mostly lies in Virginia.

Your proposal takes some pressure off disparate local communities so that they no longer need to try so hard to coexist in name=*, but it doesn’t solve the problem that semicolons are intended to solve. On the bright side, you’re using semicolons to separate language codes.

aseigo:

requires changes to be made to all name tags for it to be consistent. Other than being a fair amount of work, even if mostly automated, it makes situations where the policies change very hard to track (see the Haida Gwaii example in the proposal)

I’m not entirely sure where this is coming from. My suggestion in this thread was to introduce a semicolon delimiter as an option where name=* is already set to multiple values. If any government policy causes widespread changes on the ground, then OSM probably should see widespread changes. That goes for languages, currencies, speed limits, access restrictions, parking rules, addressing schemes, flags, you name it.

aseigo:

it is not truly backwards compatible. The default case would require changes to all renderers to render the results in a pretty fashion.

I think this is a difference in expectations. A variety of renderers, representing a variety of tech stacks, as well as non-renderers, have been handling delimiters in name=* for years without any fuss. After all, it’s just a little bit of string manipulation. Granted, not all renderers do this. I don’t know how realistic it is to expect “all” renderers to implement your proposal either.