While it is certainly possible for the country to have hundreds of languages, I am pretty sure it is rare for a single village to speak them all. In such cases, the granularity could go down to smaller units than a country (and if OSM is capable of storing hundreds of millions of buildings, I'm sure it can store a few thousand of language regions). Lastly, most (not all) "name" tags are written in a specific, single language. This attempt will document data, without requiring each OSM object to have that info.
We should also keep in mind usage of such data. For example, given an object with "name", "name:it" and "name:ja_rm", and the data consumer need to decide which one to use if they are rendering a map in French. With the above data, we can rely on the default language - e.g. if default=it (e.g. in Italy), we would choose name:it. If this is Japan, than we should use "ja_rm". We would also be able to establish the language of the "name", to see when to show it vs others. On Thu, Apr 19, 2018 at 3:35 AM, marc marc <[email protected]> wrote: > what's TSKM ? > > for Switzerland, we talk about it a few months ago. > The discussion lacked participants. > but some was thinking that adding official languages on the boundary > could be useful for rendering and tracking officially multilingual > places (name with multiple value is a very heavy (=bad) compromise in > practice, especially since there are several conventions depending on > the country such as the separator which for a (bad) reason of rendering > has sometimes been chosen as" -" instead of the usual ";") > > imho maybe with this schema > one language:xx=yes for each official language > or language=xx;yy;zz where xx yy zz are the official languages > but as far as I know, nothing has been done. > > It seems to me very complicated to tag the languages spoken in a place > of variable geometry. if two families speaking the same language live in > the same street, could you add this info ? > if not, how many families is enough/needed ? > It's too subjective. > > Le 19. 04. 18 à 01:43, Thilo Haug OSM a écrit : > > An example, good luck : > > 56 TSKM, 39 Languages > > https://en.wikipedia.org/wiki/Languages_of_Togo > > 41 TSKM, 4 languages : > > https://en.wikipedia.org/wiki/Switzerland > > > > (Germany : 357 TSKM) > > > > Am 18.04.2018 um 21:49 schrieb Vao Matua: > >> I would suggest that OSM is probably not the best place for this. > >> There are many countries that have many or even hundreds of > >> languages. The lines between the places where languages are commonly > >> spoken can be quite fuzzy and often do not follow any other features. > >> A year ago I was living in a place where people living there spoke 3 > >> different languages in addition to the "official" language. > >> > >> On Wed, Apr 18, 2018 at 12:41 PM, Yuri Astrakhan > >> <[email protected] <mailto:[email protected]>> wrote: > >> > >> What would be the best tags to use for mapping language regions? > >> I would like to create a map of primary languages spoken in an > >> area. This will greatly help with multilingual maps, allowing data > >> consumers to calculate which language name tags to use for which > >> locale. This will also give OSM community a much greater control > >> over such maps. > >> > >> Proposal (relations only, must have closed polygons): > >> type=language > >> primary=xx (required) > >> secondary=yy;zz;... (optional) > >> > >> A relation may span multiple countries (e.g. US and most of Canada > >> for English), or split countries (e.g. EN and FR regions in > >> Canada). In some cases, the relation will reuse country border ways. > >> > >> What do you think? > _______________________________________________ > Tagging mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/tagging >
_______________________________________________ Tagging mailing list [email protected] https://lists.openstreetmap.org/listinfo/tagging
