Alternate label language for periods about periodo-data HOT 7 CLOSED

I'm not sure what column mapped to localizedLabel, but I think it goes something like this.

The original label is "Early Bronze Age Period (Early Cypriote I)".

The localizedLabel is our editorial cleanup of the first part of that label, for display in the interface, always in English (or at least we called it label_en in the spreadsheet): "Early Bronze Age" (dropping the "Period", and moving the rest to the alternate label).

The alternateLabel is "Early Cypriote I", which we didn't make up, but which we separated from the original label which had an internal alternate (again, always in English, translated from the original label).

We always did alternate labels in English in the spreadsheet, even when we'd pulled them out of a combined period term in another language. Consider this example:

Original term=label: "Mesjetë /Bizantine" (Albanian)
Our localizedLabel (label_en in the spreadsheet): "Medieval" (translation of Albanian "Mesjete")
Our alternateLabel (label_alt in the spreadsheet): "Byzantine" (translation of Albanian "Bizantine")

Is this enough for semantic consistency, or should we also provide the alternate term as a separate value in the original language, and then as an English translation? I'd prefer the current system for simplicity, but will it make search in other languages harder?

from periodo-data.

The localizedLabel is not always in English in the dataset. For example:

"id": "p0vn2fr2ft3",
"alternateLabel": [
  "Late Mesolithic"
],
"label": "Yngre mesolitikum",
"localizedLabel": {
  "eng-latn": "Late Mesolithic",
  "swe-latn": "Yngre mesolitikum"
},
...

from periodo-data.

There is (purposefully) some redundancy here.

A period definition always has a label that is how it appeared in the source. Examples:

• Early Bronze Age Period (Early Cypriote I)
• Mesjetë e hershme

If the label is not English or it needs some cleanup/normalization, a period definition may be given one or more alternateLabels. These are always English. Examples:

• Early Bronze Age III
• Early Cypriot III
• Early Medieval

The localizedLabel is simply there to assign language tags to the various labels. It should not contain any values that are not already a label or an alternateLabel.

Looking at the data, I see a few cases where there is a 2nd alternateLabel that does not appear in localizedLabel, which I guess is a data quality bug.

from periodo-data.

Does that mean that localizedLabel should be able to have multiple entries for a single language?

from periodo-data.

(i.e. a localizedLabel value for each different alternateLabel)?

from periodo-data.

No. Now I remember why I only included at most one alternateLabel in localizedLabel, because the JSON-LD language indexing stuff will only work with one value per language tag. So it's fine. If they ad more than one alternateLabel just put the first one in localizedLabel.

from periodo-data.

My issues are resolved in #19.

from periodo-data.