Thorough review of Spanish Translation

My team is implementing Bahmni in Honduras where Spanish is spoken. As we have unofficially reviewed the Spanish translation, we have found several incorrect or less-than-perfect translations, even in translations that are marked as “reviewed”. Considering there is much to review and no way to know what has truly been properly reviewed, is there a way we can remove the “reviewed” status and start the review process over?

1 Like

Moved to #software:bahmni category. (@sommardahl, most of these fine-grained categories here refer to OpenMRS more generally. There’s one category that covers pretty much everything to do with the Bahmni distribution of OpenMRS.)

To answer your question I don’t believe this should be a problem, but I’d have to check if some aspect of the i18n install process would break if we suddenly don’t have any reviewed translations in a language where we used to have many.

(Maybe someone on the dev team has a quick answer to this.)

1 Like

Hi @sommardahl,

There seems to be a specific locale for the Honduras Spanish: es-HN. At least it is listed as supported by Java 8. So depending on the words/phrases and whether they are specific to Honduras or rather belong to the “vanilla” Spanish, they should be provided as ‘es-HN’ or just ‘es’.

That is definitely the case for slang, but there is a common spanish that should work for any Latin American country in a professional context. But, to your point, it might be interesting to add a more specific local and use the standard ES local as a starting point. We could then compare the two later on to see if one can improve the other.

Can you point me in the right direction to get that rolling?

I just looked on transifex and it seems that all the Spanish translations are _un_reviewed. (So maybe someone already went and did this before I got to it.)

Please go ahead and start the review process. If feasible it would be really nice to share a list of the changes you make, so that any Spanish-language implementations can prepare themselves for these changes when they do an upgrade. (I guess this can be as easy as downloading the _es.json files now, and then again after you’ve finished all the reviews/corrections/translations and doing a diff.)

Personally I would be hesitant to introduce an es-HN locale unless we see a real need for it. I would assume that 95+% of the product and config language should be universal, at least across Latin America.

I think you are correct about “across latin america”, but as one of my colleagues pointed out to me earlier today, Spanish in Spain is significantly different from Spanish in other parts of the world. It could be that, by changing “es” to match Latin America, we alienate European Spanish speakers.

We spoke this morning and we decided it will be a better approach to deploy Bahmni with the “es” locale and let the users tell us what to change based on what they see. So, we will forego a system-wide translation/review effort up-front in lieu of a more agile approach. We will either be correcting, translating and/or reviewing along the way.

This is true, but since Bahmni targets low-resource settings, we’re much more interested in the Latin America use case. I would make Latin American Spanish the Bahmni default.

This makes sense.

One thing to note is that Transifex lets you add comments to translations, so if people are in there translating, and they come across something suspect, they can add a comment for later review.

So you know the history, the initial Spanish translation was done by a pair of Ecuadorian developers on the Bahmni team (so, fluent in computer Spanish, but not medical Spanish). And I believe that further additions were made for an endTB deployment in Peru.

I am following up on this thread from last year about Spanish translations for use in Central America. Were you able to finish the translation and deploy OpenMRS/Bahmni? Thanks Hamish