There are two possible approaches:
- You clearly separate one variant from the other, i.e. treating them as completely different languages, or:
- You accept that most terms are the same in all variants, and you only mark the regional variants explicitly where necessary.
When managing termbase languages in MultiTerm, the consequences of each approach are as follows:
Approach 1 - Separate Indexes
- In your termbase definition, you add different indexes per language variant, i.e. English (United States), English (United Kingdom), English (Australia) etc.
- You do not treat English terms within one entry as synonyms, but rather as "translations" of each other (subway to be the American translation for British underground)
- None of the language variants is treated as the "main" or "parent" variant.
- For a complete coverage of language variants, you need to fill in all terms in all language variants. This can require some work and extra effort.
Approach 1 is useful if your termbase contains only a few entries that have no variants, and most of the entries always contain variants. This is a rare scenario.
Approach 2 - One Index
- In your termbase definition, you add only one, neutrally named, index for all languages belonging to one "parent"; for instance English.
- You create a picklist value field, for instance called Regional Variants, with values like United Kingdom, United States, Australia etc., maybe also including other regions such as Germany for other language variants you plan to cover (all regions in one picklist). This picklist field should be a child field at term level.
- You select this picklist field only when required, i.e. for cases like subway/underground, while not specifying anything further for terms that are the same in all variants.
Approach 2 is probably the more pragmatic and cleaner approach and therefore recommended for most cases.