[Talk-in] Multilingual name tagging convention in India

Anish Mangal anishmg at umich.edu
Thu Jul 21 15:30:08 UTC 2016


Interesting discussion. Currently it seems the name tag by default may
not be English in some places (in India) and the name:en tag may not
exist.

eg: http://home.braddock.com:28112/osm/slippymap.html

The ordering of labels was
COALESCE(tags->'name:en', tags->'name:hi', tags->'int_name', name)

So, in some cases name:en doesn't seem to exist (or maybe I'm wrong?)

On 7/21/16, Arun Ganesh <arun.planemad at gmail.com> wrote:
> Some time back, I did some quick research on how we officially recorded the
> names of places in India, might be an interesting read:
> https://gist.github.com/planemad/decdf608776726de1007
>
> Related projects that could help expand language coverage in India is
> attempting latin to indic language and cross indic language
> transliterations:
> - https://github.com/wikimedia/jquery.ime/tree/master/rules
> - http://libindic.org
>
> On Thu, Jul 21, 2016 at 12:01 PM, Arun Ganesh <arun.planemad at gmail.com>
> wrote:
>
>> Its surprising we all agree that using the latin script as default for
>> local names throughout the country is most usable. Wish we had more
>> perspectives from rural users of the map, but I guess at the moment we
>> have
>> not reached there.
>>
>> This issue would come up again if that happens and local scripts are used
>> for the default over latin, something similar is already happening in
>> Bangladesh http://www.openstreetmap.org/#map=11/23.1262/89.1843
>>
>> For now, I guess lets freeze on the latin alphabet as default for all
>> urban areas of the map. To sum up:
>>
>> - The default `name` of a feature is that name and spelling that is
>> recognized by the local public, preferably recorded in the latin script
>> for
>> data consistency throughout the country. Use a localized script only if
>> there is no recognized latin spelling for the name
>> - The localized spellings and scripts should be added in the
>> corresponding
>> `name:lang` tags
>> - Due to complications arising over the ordering of multiple languages
>> and
>> scripts in a label, no `name` tag should contain more than one variation
>> of
>> the name, use a different tag for each variation
>>
>> If no major corrections, I can add it to the wiki this week and we can
>> make changes there.
>>
>> On Thu, Jul 14, 2016 at 3:59 PM, Arun Ganesh <arun.planemad at gmail.com>
>> wrote:
>>
>>> Moved some old instructions for setting up a tileserver that Johnson
>>> wrote to https://github.com/osm-in/tileserver
>>>
>>> Anyone interested in setting up a custom tileserver with Indic language
>>> rendering, please feel free to expand.
>>>
>>> On Thu, Jul 14, 2016 at 12:16 AM, Chetan H A <chetanha91 at gmail.com>
>>> wrote:
>>>
>>>> I totally agree to follow current convention we have now. Multilingual
>>>> names can't be supported in India. We have 22 official languages in
>>>> India
>>>> and If every language names were added with English names will clutter
>>>> the
>>>> map. It's good to add separately using " name:lang=* " tag.
>>>> http://taginfo.openstreetmap.in/search?q=name%3A
>>>>
>>>> Best,
>>>> Chethan
>>>>
>>>> On Tue, Jul 12, 2016 at 4:23 PM, Arun Ganesh <arun.planemad at gmail.com>
>>>> wrote:
>>>>
>>>>> Not having map tiles available for the various languages is definitely
>>>>> a problem, which motivates one to tag in such a way just so that the
>>>>> local
>>>>> name is visible. Efforts to create an Indic language tileserver have
>>>>> not
>>>>> really taken off, and probably this is what we should reconsider with
>>>>> more
>>>>> seriousness.
>>>>>
>>>>> Experiments so far:
>>>>> - http://openstreetmap.in/demo.html
>>>>> <http://openstreetmap.in/demo.html#9.99/13.9372/75.6525> which is a
>>>>> little futuristic but very flexible for as many languages as possible.
>>>>> One
>>>>> blocker is that complex text shaping to render Indic scripts correctly
>>>>> is
>>>>> unknown when this will happen
>>>>> https://github.com/osm-in/openstreetmap.in/issues/14#issuecomment-193619651
>>>>> - http://mlm.jochentopf.com looks like a good experiment worth
>>>>> replicating, but there are not much details on how it was done
>>>>> - http://yogiks.github.io/osm-kn/map/ World Kannada map made by
>>>>> Yogesh. At the moment probably the most practical approach.
>>>>>
>>>>> If we have a regularly updated working demo of Indic language tiles
>>>>> for
>>>>> the major languages, this could better demonstrate the practical
>>>>> application of using the name tags correctly. Anyone have other
>>>>> experiments
>>>>> to share?
>>>>>
>>>>> On Tue, Jul 12, 2016 at 2:14 PM, Yogesh K S <yogesh at itforchange.net>
>>>>> wrote:
>>>>>
>>>>>> I too agree with keeping the current convention of name tags for
>>>>>> multilingual names. Belgium has just two languages and in context of
>>>>>> our
>>>>>> regional languages, first mapper rule can create conflicts as Arun
>>>>>> indicated previously. Perhaps we should try to improve the current
>>>>>> state of
>>>>>> multilingual names in India -
>>>>>> http://taginfo.openstreetmap.in/search?q=name%3A
>>>>>>
>>>>>> cheers,
>>>>>> yogi
>>>>>>
>>>>>>
>>>>>> On 07/05/2016 06:19 PM, I Chengappa wrote:
>>>>>>
>>>>>> It would help greatly if mappers could see names in other languages
>>>>>> on
>>>>>> a rendered map. I see that the demo, for instance at
>>>>>> http://openstreetmap.in/demo.html#9.99/13.9372/75.6525 is still just
>>>>>> a demo.
>>>>>>
>>>>>> On 5 July 2016 at 07:19, Srravya C <srravya.c at gmail.com> wrote:
>>>>>>
>>>>>>>
>>>>>>> Found a little more about this topic in a discussion here
>>>>>>> <http://wiki.openstreetmap.org/wiki/Talk:WikiProject_India#Use_English_for_the_.22name.22_tag>.
>>>>>>> We could put the details about multilingual naming for India in this
>>>>>>> page <http://wiki.openstreetmap.org/wiki/Multilingual_names>. So
>>>>>>> that anybody referring to this page in the future for naming
>>>>>>> convention,
>>>>>>> will get an idea about existing conventions followed for India.
>>>>>>>
>>>>>>> On Tue, Jul 5, 2016 at 8:44 AM, Thejesh GN < <i at thejeshgn.com>
>>>>>>> i at thejeshgn.com> wrote:
>>>>>>>
>>>>>>>> I think the existing convention works really well. Use name tag for
>>>>>>>> latin script and specific language tag for that language names. I
>>>>>>>> think
>>>>>>>> this makes it easy to work and also divide work.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Jul 4, 2016 11:34 PM, "Arun Ganesh" <arun.planemad at gmail.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> Bringing this up since there are recent edits adding multilingual
>>>>>>>>> labels to the `name` tag [1][2].
>>>>>>>>>
>>>>>>>>> The current convention that has been followed is to use `name` for
>>>>>>>>> the local name in latin script. The reasoning if I remember was two
>>>>>>>>> fold:
>>>>>>>>> - English being one of the official languages, latin script is
>>>>>>>>> widely used in cities and in public signages in smaller towns. Only
>>>>>>>>> in
>>>>>>>>> villages it is uncommon.
>>>>>>>>> - Keeping the tagging convention simple[3], that could be used
>>>>>>>>> throughout the country in a standard way
>>>>>>>>>
>>>>>>>>> Keeping just one name in each tag, we have been able to avoid
>>>>>>>>> issues like the first mapper rule[4] that could cause unnecessary
>>>>>>>>> regional
>>>>>>>>> conflicts over ordering of scripts like in Belgium (and they have
>>>>>>>>> just two
>>>>>>>>> languages).
>>>>>>>>>
>>>>>>>>> Since we can add the regional names in their own name:lang tag, no
>>>>>>>>> data is ultimately lost, and it is upto the data consumer to choose
>>>>>>>>> which
>>>>>>>>> language to support for the user.
>>>>>>>>>
>>>>>>>>> Would be great to hear positives and negatives of this approach
>>>>>>>>> and
>>>>>>>>> how it benefits the public Is it time to revise this convention to
>>>>>>>>> be more
>>>>>>>>> regional rather than national?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> [1] http://www.openstreetmap.org/changeset/40472998
>>>>>>>>> [2] http://www.openstreetmap.org/changeset/40479683
>>>>>>>>> [3]
>>>>>>>>> http://wiki.openstreetmap.org/wiki/WikiProject_India#Naming_in_different_scripts_and_languages
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Arun Ganesh
>>>>>>>>> @planemad
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> Talk-in mailing list
>>>>>>>>> Talk-in at openstreetmap.org
>>>>>>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>>>>>>
>>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Talk-in mailing list
>>>>>>>> Talk-in at openstreetmap.org
>>>>>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Talk-in mailing list
>>>>>>> Talk-in at openstreetmap.org
>>>>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Talk-in mailing
>>>>>> listTalk-in at openstreetmap.orghttps://lists.openstreetmap.org/listinfo/talk-in
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Talk-in mailing list
>>>>>> Talk-in at openstreetmap.org
>>>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Arun Ganesh
>>>>> @planemad
>>>>> <http://j.mp/ArunGanesh>
>>>>>
>>>>> _______________________________________________
>>>>> Talk-in mailing list
>>>>> Talk-in at openstreetmap.org
>>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Talk-in mailing list
>>>> Talk-in at openstreetmap.org
>>>> https://lists.openstreetmap.org/listinfo/talk-in
>>>>
>>>>
>>>
>>>
>>> --
>>> Arun Ganesh
>>> @planemad
>>> <http://j.mp/ArunGanesh>
>>>
>>
>>
>>
>> --
>> Arun Ganesh
>> @planemad
>> <http://j.mp/ArunGanesh>
>>
>
>
>
> --
> Arun Ganesh
> @planemad
> <http://j.mp/ArunGanesh>
>


-- 
Anish



More information about the Talk-in mailing list