Is there specific design purpose for the local folder of lu file being language code only vs full LCID string?

[dawwa]

Wondering if there is case we concern about the full LCID string, e.g., "us" part of "en-us"

So far, all the code seems only concerning the language code.

For example:

• The *.lu files container folder is "en" instead of "en-us". https://github.com/Microsoft/AI/tree/master/solutions/Virtual-Assistant/src/csharp/assistant/CognitiveModels/LUIS/en

• And the tool scripts seems always only concerning about the language code. https://github.com/Microsoft/AI/blob/master/solutions/Virtual-Assistant/src/csharp/assistant/DeploymentScripts/generate_deployment_scripts.ps1 $langCode = ($locale -split "-")[0]

• The bot file generated also only uses language code part. https://github.com/Microsoft/AI/blob/master/solutions/Virtual-Assistant/src/csharp/assistant/DeploymentScripts/update_published_models.ps1 $langCode = $botFileName.Substring($botFileName.Length - 2, 2)

[lauren-mills]

Hi @dawwa, the main reason for using only the language code is for cases when different locales map to the same language. For instance, en-gb and en-us would both be using the same English services and responses in our current implementation. That being said, for a more complicated scenario you could use the full string to have different behavior/responses depending on the country code. It’s simpler at this point to use one configuration per language.