add a language property

[p1d1d1]

For some datasets in our catalog, we provide asset files in different languages. In our current implementation (e.g. https://radiantearth.github.io/stac-browser/#/external/data.geo.admin.ch/api/stac/v0.9/collections/ch.bav.haltestellen-oev/items/haltestellen-oev?.asset=asset-haltestellen-oev_2056_de.csv.zip) we are using an extension to provide the "lang" info at asset level.
By the way we think that the file extension would be the right place for that, something like "file:lang".
What do you think? Would you consider to add "lang" to the file extension?

[m-mohr]

Charset for textfiles would be a (text) file attribute, but I don't think the language itself is really a "file" attribute so I think language could be a separate extension itself. This could simply be based on what OGC APIs define: http://docs.ogc.org/DRAFTS/20-004.html#collections-properties-table (i.e. a language property). Long-term I could see this also being added to STAC common metadata, honestly. Maybe it would be worth opening an issue in stac-spec itself about the addition of language to common metadata (I'd support that).

PS: I'm likely working on a "multilanguage" STAC API extension soon so there might be something coming out of that (although in an HTTP context usually you do it via headers): radiantearth/stac-api-spec#306

[p1d1d1]

@m-mohr as far as I can see, the language property if for the metadata (title, description, ...).
Here we need something really at file level. We want to communicate to the user the language in which any textual information in the file is expressed.
PS: we use STAC also for vector data !

[m-mohr]

@p1d1d1 If we'd add "language" to the STAC common metadata it would be applicable to different contexts, comparable to "created" and updated" for example:

• language in Item properties / Collection -> metadata language
• language in assets -> file language
• language in links -> language of what is linked to
• ...

PS: I like the use of vector in STAC. Is there something we can look at?

[p1d1d1]

@m-mohr

If we'd add "language" to the STAC common metadata it would be applicable to different contexts, comparable to "created" and updated" for example:

Ahahh, now I see! Then I'll open an issue on stac-spec

PS: I like the use of vector in STAC. Is there something we can look at?

https://radiantearth.github.io/stac-browser/#/external/data.geo.admin.ch/api/stac/v0.9/. Most collections are about vector datasets

[m-mohr]

I've also mentioned it here: radiantearth/stac-spec#1187