How to handle translations with translation bureaus?

Question

I'm wondering what's the best way to realise translations with translation bureaus.

Basically I see two ways of doing this:

1. Send them the CSV files.
2. Give them access to a Magento installation for inline translation.

The issues that can occur:

1. CSV files

   • Quite many agencies prefer "professional" file formats like .pot / .po (GetText).
   • The context for the strings to be translated is missing.
   • Not every text can be found in CSV files (CMS block, CMS pages, ...).

2. inline translation

   • Not every agency wants to learn how to use the Magento backend
   • It's difficult to reproduce all scenarios for translatable texts (think error messages e.g.)
   • I want my translations to be in versionable file instead of the database. (Ok, you can work around that and extract the translations from the DB to files.)

All in all this seems to be time-consuming and error-prone. Am I missing something obvious here?

Answer 1

To me, best way of this giving access to inline translation. As you might guess, perfect translation should be what you see what you translate. For instance, in Turkish translate, many common words has might be different meanings. Therefore, translator should see which part of the Magento translating.

Answer 2

It's a difficult question that we meet all of us. I give the CSV files (or convert to po if they want). As you said with inline translation, you cannot cover all scenarios, it's time consuming, error message are most of time missed. The solution which provide less problem for me is CSV, I use https://github.com/diglin/Diglin_LanguageCsv to extract them from a module.

However, regarding CMS pages for example, it may be better to give access to the inline translation. By defining the set of pages to be translated, it should not be a big constrain for translators if you are clear with them.

If you work regularly with the same agency and translators, they and their Memory manager tool will get used to the "shop vocabulary".

So depending of the project, I see a mix of both solutions. In my case, I provide mostly CSV file.