Transliteration changes

I noticed that Sundar1 has been changing the transliteration of name=* values in many places across India, e.g. changing “Chowk” to “Chok” -

They responded to my changeset comment with a detailed explanation in PM, and I figured we had best discuss the matter with the wider OSM community, especially Indian mappers.

Hello, I’m the culprit and I’d like to explain myself.
Firstly, thank you to contrapunctus for introducing me to this forum. I’m very much looking forward to exchange views with Indian contributors to OSM.
Like Wikipedia, to which I’ve been a long-time contributor, I find OSM a very intriguing project which I rather quite recently joined and I’m far from familiar with. Yet, due to the nature of Wikipedia and OSM, I have quickly noticed very similar problems.
I see a precondition for a fruitful discussion in an agreement on the self-image of OSM. Apart from being a bottom-up project, like Wikipedia, I would think that OSM aims to cater to users in the whole world just like google maps, bing, etc. and thus needs to be a standardised map adhering to a set of common principles and rules. This is my guiding principle when working on OSM as well as on Wikipedia. And this is not to say that, e. g. google or bing, are standarised when it comes to spelling rules - it’s quite the contrary.
So, what standards or rules am I talking about? As for unified mapping symbols that are used in all countries there should also be a unified spelling system similarily recognisable in all countries. Do we take into account an international audience and apply such standards or do we blindly copy into our map whatever we happen to find along the roads? What would be more logical than to use the same system as the Survey of India despite the known shortcomings?


You might find interesting.

We don’t try to impose order where there is none. If the street is signposted with a particular name, that’s what we map. It’s not our role to correct unusual orthography, no matter how alien it is to use. See the “Original research always wins” section of that page on the OSM wiki.

Maybe, but with the best will in the world, we probably don’t want to re-examine the entire basis of the project on the viewpoint of someone who’s been here for a few days. :slight_smile:

Sundar1, can you point us toward the “same system as the Survey of India”?


I expected a lot of more interest in this topic, especially from Hindi native speakers and the little input is quite disappointing. Alas, more might come …
At least I want to answer to the few responses.

Of course, you do! Everyone is required to use a whole set of symbols for mapping supplied by OSM. And, of course there is an order for spelling. Why not apply it? I’m not talking about changing any names on the ground, I’m talking about transiterating local script into Roman script.
Map users will find it pretty confusing to see a variety of different spellings for the same name even within the same location as a result of using transliterations that people use “on the ground”. As a matter of principle, transliteration must always be based on the local language/script which in our case is Hindi/Devanagari. So the question is, should people on the ground, more or less familiar in writing with the Roman alphabet, transliterate Devanagari into Roman letters any way they see fit or should they get some guidelines (“imposed order” in your words)?

My question about the transliteration is based on the assumption that OSM, calling itself a map, most of all, wants to be as good a map as any other if not better, just as legible, informative and useful. A precondition for any map is standardisation. For that purpose OSM uses unified symbols which, after all, are also “imposed”. The fact that OSM also is “open” doesn’t change it’s main purpose of being a map. If there is no agreement on this, we need not discuss any further and I’m in the wrong place.

Who wants to re-examine the entire basis - why this exaggeration? This is only about the transliteration of a certain language. Using the royal “we” in "We don’t try to impose … " is very telling. And being dispatched to the back of an imagined queue for beginners who’ve “been here for a few days” is not well made up for by a smiley face at the end, especially to someone who has worked with maps and mapping for 50 years. I beg to answer why my viepoint on transliteration would rock the entire basis of OSM? Even if I’m a beginner in OSM, I consider it quite legitimate to discuss a unified transliteration within a map as it concerns all mapping. I also consider a native English speaker pre-disposed. Perhaps you’d like to focus on the issue I presented and give me your qualified -and if you’re into mapping- professional opinion.

indigomc, you at least ask a question.

To find references to the Survey of India is not easy. I’m familiar with Indian Survey maps because I used them during my time in Nepal.
I suggest you google

–The United Nations and Standardization of Geographical Names
–Standardization of Geographical Names in India
–The Romanization of Toponyms in the Countries of South Asia

There’s also a lot about transliteration conventions on Wikipedia where, of course, spelling is also an important issue.

quite a recent book: Mapping Place Names in India by Anu Kapur

The general pattern has been to use the common name, ‘on the ground’, that which is used in general practice. Many people take Wikipedia usage as a guide. I prefer to transliterate for all less common names.

A problem with other systems is that they are incomplete or ‘incompetent’, they do not cover all phonemes in the differing languages. Possibly in any languages, they are still based on western languages.