[OSM-talk] nomoj de internaciaj objektoj / nazwy obiektów międzynarodowych ? names of international objects

Martin Koppenhoefer dieterdreist at gmail.com
Fri Dec 6 08:55:34 UTC 2019


Am Fr., 6. Dez. 2019 um 08:08 Uhr schrieb Maarten Deen <mdeen at xs4all.nl>:

> On 2019-12-05 22:12, Tomek wrote:
> The adjacent Caspian sea is displayed as دریاچه خزر which is not the
> complete name tag (Каспийское море / Хәзәр дәнизи / دریاچه خزر) but
> looks to me the name:fa (Farsi).
> For me, that is very bad since I can not read Farsi so I don't know what
> is written there. Same applies to Chinese, Japanese, Korean, and a whole
> host of languages that do not use the latin script.



it can be done (and is already done in some context) to display names in
your own language (or another fallback) in case there aren't letters in
latin script (even an automatic transliteration at display time may be
better than not being able to read the script at all).


IMHO Openstreetmap does try to be politically and nation independent.
>
> Still, it is a Good Idea to have one standard (language) to communicate
> or define things, like everything meant for an international public in
> the wiki is English and the tag system is English.
>


you should argue why it is a good idea to have _one_ standard language in
the project. IMHO it excludes many billions of people from participating,
and confirms the Anglo-Saxon dominance in the tech world also in  the
mapping world. It is convenient for us Europeans, because most of us are
able to communicate in English, but I am not sure we could not do better at
integrating people from different cultural contexts. It is also a sign we
are sending out to the others, whether there is "the main language
English", or whether we communicate that every language is accepted. In
practical terms, I agree it is hard to imagine how you and me could
contribute to a proposal in Chinese, Farsi or even Swahili. Also reducing
the "any language" to "the UN languages" would bring in serious (currently
unsurmountable) overhead if we required for instance proposals to be
translated into all UN languages.




>
> > Suggestion 2b: use the name in a neutral language, i.e. planned or
> > extinct: Lingvo Internacia Esperanto (EO), Interlingua (IA), Ido (IO),
> > Latin (LA), I don't know Latin, so I would need help.
>


none of these is "neutral" and more importantly, none of these seems
"inclusive" and suitable to broaden participation, they are all either
elitist, or at least spoken by very little people, and likely both. From a
practical point of view, most people worldwide are able to communicate in
English and Mandarin Chinese, according to SIL International, 2019, there's
just a 20 Million difference in favor of English speakers, and there is
three times more people with Chinese as mothertongue). Following in this
list [1], there is with roughly half the speakers, Hindi and Spanish. Again
half of these, French, Arabic, Bengali, Russian, then Portugese,
Indonesian, then Urdu, then German and Japanese, followed by Swahili as the
first language below 100 Million speakers, and so on.



> > Suggestion 3: for a place with fewer names, I suggest using several
> > labels:
> > wikipedia: ru = Чёрное море + wikipedia: ro = Marea Neagră,
>
>

What have wikipedia article links to do with "labels"?



> > for other places:
> > Suggestion 4a: remove the label "wikipedia" and leave only "wikidata",
>
> I agree that it is strange to have (e.g. for the Caspian Sea
> multipolygon 3987743) wikipedia=en:Caspian Sea when it is not in
> England. Is there a reason for that other than historic? Since there is
> also a wikidata link.
>


I am strongly opposing the idea of removing wikipedia article links when
there are wikidata feature links. The former are human readable and mostly
(?) the original data that the mapper added, the latter are often
automatically derived (from wikipedia article references), worse verified,
link to a less mature project (a wikidata entity to which I link today may
change its nature significantly until tomorrow), are not human readable and
a typo in just one digit makes them completely worseless, and there is no
1:1 relation of wikipedia articles and wikidata objects (this is btw. a
serious problem which I do not know why the wikidata community doesn't
address, they seem to assume that there is just one wikipedia article for
one wikidata object and vice versa).

Please also note that "wikipedia:en" is not about "England", it is the
knowledge of the world collected in the English language (btw. it is the
wikipedia version with most articles). With all



>
> > Suggestion 4b: add links in a neutral language: wikipedia=ia: Mar
> > Nigre
>
> Why is Anguilla a neutral language? Mar Nigre looks French to me, why is
> that a neutral language? Should it be Esperanto? Why would that be a
> neutral language since it is
> written in latin alphabet? Also Esperanto to me seems more like a
> western language than an eastern/asian language.



it is clearly a Roman language (intended as somehow strongly related
to/derived from Latin), as is Esperanto and probably many more (e.g.
interlingua sounds so as well). None of these can be considered "neutral".
IMHO we do not gain anything from using a language that "nobody speaks"
(compared in numbers with natural languages). While it is probably
undisputed that we should add name:language tags as much as possible, I
also liked the idea of putting the United Nations languages in the "name"
tag for big "international" features (like the north pole, antarctica, the
atlantic and pacific ocean, etc.), while it seems to make sense to have
just the locally spoken adjacent languages for smaller features (e.g. the
channel between Britain and France is currently tagged in French and
English and this seems sensible).

> Suggestion 4c: add more links, but in which languages?

if "links" is about wikipedia article references, there is already the rule
that one link is sufficient where the other language article is linked as
corresponding in wikipedia. If this is not the case, the suggestion is to
add wikipedia:languagecode=* tags.
I agree that this convention is a compromise to save on wikipedia links
(and retain readability) at the cost the issues may arise from people
changing the structure of wikipedia, which leads to different articles
being accessed through interlanguage links as those that had been in place
when the mapper checked.


Cheers,
Martin


[1]
https://en.wikipedia.org/wiki/List_of_languages_by_total_number_of_speakers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20191206/85e5f588/attachment.html>


More information about the talk mailing list