Topic on Talk:Map improvements 2018/Flow

Great to see the publicity around this project.

OSM has a long-standing community consensus that transliteration does not belong in the OSM database: https://wiki.openstreetmap.org/wiki/Names#Avoid_transliteration . There are concerns that this project could encourage people to add transliterated names.

Could you clarify your instructional/publicity materials to emphasise this?

Alternatives are: adding a Wikidata id to the OSM object, so it can be crossreferenced to a set of transliterated names in Wikidata; or machine-generated on-the-fly transliteration (this has been successfully carried out by the German OSM community for many years now).

Hey IP, could you help me understand the concerns a little better?

We've been telling folks that if there isn't a label in their local language on Wikimedia Maps to go to OSM to add a label in their local language (like "name:fr=Foo"). No encouragement in either direction regarding bulk automation. Which, if I'm reading the linked article correctly is the largest concern? How might we better update our documentation here on MW.org to make it more apparent that transliteration is a nuanced thing in the OSM community? It's a wiki, feel free to edit directly or make a suggestion here. :)

The Wikimedia Foundation is not doing any active promotion of maps as the project is winding down and entering ongoing maintenance (no new features for the time being). In fact, in short time this page will be marked as historical and ongoing discussions will encouraged to take place on Wikimedia Maps' talk page. I say this to alleviate any immediate feelings of concern and suggest that we perhaps work on adding documentation to that page going forward.

Edit: I just now saw the comments on the blog post. That provides more context, but I am still happy to have more feedback.

The comments on the post put it well. Thanks for your understanding.

My personal opinion is that data should be useful. OpenStreetMap is open source and I am totally in favour to have all the names available as they are known in any language. This allows us to provide maps we need in all our Wikipedias. When Wikidata and OpenStreetMap are linked on the object level, then it is irrelevant where useful labels are kept. You can avoid transliteration as much as you like and we still have the functionality that is required. That is the beauty of collaboration and that is the beauty of open source. Thanks

Hi Gerard, that's fine, but your personal opinion doesn't trump OSM best practice when it comes to entering data to OSM. It's very likely that OSM contributors will revert widespread additions of transliterated names. Encouraging people to add data which will only be reverted doesn't help anyone.

OSM wouldn't be so irresponsible as to ask thousands of its contributors to break Wikipedia rules. Please don't encourage thousands of your contributors to break OSM rules. Thank you.

Hoi, the point of this best practice is that it served you well. You want at least a situation that is as good as you have it now. What you have to consider is what use Wikimedia will make of your data. It is likely to become the single biggest usage of OSM. What I want is to consider options, see how we can make things work. On your servers on ours, the point should be that we play nicely together. I am not interested in a public conversation with you, I am interested in talking about the needs that I perceive. When you care to hear from me as much as I like to hear from you, I am sure you will contact me.

> It is likely to become the single biggest usage of OSM

Given that Facebook, Apple, Snapchat, and many others use OSM, I doubt it. Please have a little more humility and respect for the settled opinion of the OSM community rather than trying to bully us with your self-perceived importance.

Gerard, did you read the link that the IP user included in their first post? Here it is again: osmwiki:names#Avoid transliteration. The most relevant part is:

As well, many transliterated names frequently have been imported as is from Wikipedia (or Wikidata) for naming their articles, but the names may have been chosen quite arbitrarily on these wikis. It's not needed to import these transliterated names in OSM: we can just link to a single Wikidata entry or a single Wikipedia article in a single language, preferably the main language used locally, in order to find the other articles.

So data *is* "being useful": the purely geographic and *on-the-ground verifiable* data is being useful in the OSM database, and the encyclopedic and otherwise cultural relations are being useful in Wikidata.

Arlo Barnes, you are not saying anything that was not said before. You do not add a thing and thereby make things worse for your point. Talk to me privately, not publicly is what I said. I will not engage in any argument in this way as it is not productive. I only get a repetition that does not improve the argument.