In our web site we use .resx files to provide labels and GUI in 4 different languages (English, Spanish, French and German). A specific department is in charge to provide translations, given the english values (default). We, as programmers, define english translations and send them to the language department in a specific day. Usually after a week we get the translated list back and we integrate it in the solution.
However between the date we send the list out and we get it translated, it might happen that new labels are created (usually between 10 and 20 entries) and managed internally by another department to save time.
What would be the best practice to manage and process the "delta" entries that need to be translated and then integrated in the labels list?
How current approach is try to sort the .resx files and then compare them to find out the new fields missing a translation. But I guess there is a better approach for this.
I'm working for the Translation Agency Supertext and we often have this problem too. But most translation departments use tools like Trados, MemoQ,etc. They all have built in Translation Memories. E.g. everything that has already been translated once, can automatically be translated again.
In your case, you should be able to just send them the files again and they can just add the missing translations. If your department is not using any tools, I would have a serious word with them...
Alternative there are online tools like transifex, where you could actually manage this process yourself. There are other similar tools, but would have to dig them up.
In any case, the core of the message is that you use a TM (Translation Memory).