In Thailand, it is customary to name most objects using the only official language, which is Thai. To adhere to the global recommendation, mappers have been adding Thai names by duplicating values using both the name and name:th tags. However, over time, these tags have been subject to individual changes by inexperienced mappers, leading to inconsistency in the data.
Given the small size of the regular contributor community, maintaining consistency over time has been challenging. As such, I’m wondering whether it would be simpler to remove the name:th tag by default when the name tag is already in Thai language. This would help to reduce confusion as other local languages typically used for POIs, such as English or Burmese, use an alternate alphabet.
I would appreciate your suggestions on this matter.
if object has only name in a single language then having also for example name:th is less useful
For example in Poland typical shop name will have only name without name:pl
Repeating name in language specific tags is far more useful when object has names in multiple languages (as otherwise more advanced labelling with fallback to other languages becomes impossible)
But anyway, the recommendations are what they are and despite the downside of potentially having conflicts in the duplicated data, I agree it is better to do it that way. First of all, the conflicts can be found automatically and could also be fixed semi-automatically, so this is not that much of an issue.
From my own mapping experience in Thailand and Myanmar, I know that having all three of name, name:en and name:th is better exactly for the case that inexperienced mappers update data: Oftentimes only the English name was recorded first (by tourists or whatever), into name. Then, as locals or people who would get a Thai/Burmese keyboard joined in contributing, oftentimes what happened was that name was simply overwritten by the non-transliterated name. This is not good. This is why if the name in any one language in regions where street signs commonly have names in several languages and/or in the country’s language plus transliteration should also be put into name:<lang> so that it is preserved when people “fight” (inadvertently or deliberately) about what should be in the name tag.
Also, tagging all three names helps renderers and other software to recognize which language the name is in, which is important for localization.
I’m familiar with the existing Thai guidelines, I was just looking to get some feedback from a global perspective
Can you clarify what you mean by “semi-automatically”? As far as I know, when both name and name:th are in the Thai language and have conflicting values, it’s not possible to determine the most up-to-date or accurate value through automated means. Each object’s history needs to be manually reviewed to resolve the conflict.
Well, it is however possible to find which name was changed last. Semi-automatically because one needs to search through the object’s history and there is no powerful API for that. I.e. for every offender, one has to do several API calls to puzzle together the history of the element. This makes it more cumbersome to work with because I think there are not many tools that can do something like that out of the box. E.g. Overpass as far as I know not.
