Description of problem: In some Zanata projects we will have the same English string repeated multiple times, which may lead to different translations for the same string. While some times this may be intentional due to context differences, sometimes this may be an error. A possible solution from Christophe (cfergeau): "This looks like something which would not be that hard with the appropriate tools. Zanata (or some separate tool) could have a view only showing strings (+ translations) which are duplicated, translators would be able to review these strings, and could easily spot the source strings which are needlessly duplicated. Once you have a list of strings which should have a common translation, this can be given to the developers and fixed."
This should not be difficult to solve now that Admin and User Portal strings have been merged into a single file, oVirt (master). Could we extract these duplicate strings in an automated fashion please?
Another way this could be solved in Zanata is by auto-filling duplicate translations once one has been translated, referencing the original translation. That way it can always be reviewed later in case it is in a different context. Another option could be to group them together under one translation. This will always be better done by developers, but a lot of times they will miss this, so something in Zanata will help. Would these options solve the problem?
'auto-filling duplicate translations once one has been translated' is what is currently used by other translation softwares like SDLX and MemoQ. You may which to auto-fill but save them as Fuzzies to allow context discrepancy control.
Migrated; check JIRA for bug status: http://zanata.atlassian.net/browse/ZNTA-78