Evaluate potential for localisation of hubs

[annakrystalli]

There's been discussions on the potential for localising hub infrastructure recently.

I think the two main areas that would need translating are:

1. The columns in mode output data files
2. The std task ID properties in schema.

Accessing localised versions of std model output colnames

Given we've recently started exporting objects of std hub terms (e.g. std_colnames) and given a recommended option for accessing data through a package is through a dedicated function a possible option for localising std colnames is by:

• turning std_colnames into a matrix, one row per language of translated terms.
• Use a function that takes term names as an argument and the required language as the second to return the values in the language required:

library(hubUtils)

# Create matrix of localised terms
std_colnames_m <- std_colnames |> as.matrix() |> t()
std_colnames_m <- rbind(
    std_colnames_m,
    c("modelo_id", "tipo_valore", "id_valore", "valore")
)
row.names(std_colnames_m) <- c("en", "it")
std_colnames_m
#>    model_id    output_type   output_type_id   value   
#> en "model_id"  "output_type" "output_type_id" "value" 
#> it "modelo_id" "tipo_valore" "id_valore"      "valore"

std_colnames <- function(terms, lang = "en") {
    std_colnames_m[lang, terms]
}

# Access localised version of terms
std_colnames("model_id")
#> [1] "model_id"

std_colnames(c("output_type", "output_type_id", "value"))
#>      output_type   output_type_id            value 
#>    "output_type" "output_type_id"          "value"
std_colnames(c("output_type", "output_type_id", "value"), "it")
#>    output_type output_type_id          value 
#>  "tipo_valore"    "id_valore"       "valore"

^{Created on 2023-07-17 with reprex v2.0.2}

[nickreich]

Noting that in another discussion about this request, it arose that perhaps we could broaden the scope of this issue to not just address "localization"/"internationalization" but also if a hub just wanted to use names for std_colnames that were more intuitive for their group. E.g., could a hub decide that a better set of names might be:
(value_type, value_type_idx, value)
Could a solution to the "localization" problem also introduce the ability for teams (or the hubverse org) to specify "allowable" sets of std_colnames?