language abstraction

[rjoberon]

We need to figure out how to handle contexts within different languages. Currently, this is solved by having separate context files for each language and attaching the language code to the context file's name.

[perasperaadastra]

I would argue that language abstraction is not necessary and all elements of the object and attribute sets should be seen as arbitrary strings.

Additionally, we already have the case that the two versions of bodies_of_water "look" the same (e.g. isomorphic contexts with just some relabeling), but actually are different datasets, as seen by the different sizes of object/attribute sets.

Thus, we should

• close this issue
• rename the German version of bodies_of_water context and handle it as its own separate context
• remove all language codes from the context names
• update readme
• discuss if we need to change the metadata schema as well