Hi Matthew, 2008/10/15 Matthew East <[EMAIL PROTECTED]>: > > The strings are not identical: they are just similar. Rosetta has no > way of remembering translations of similar English strings (for > obvious reasons). >
I have not checked whether all the strings were 100% identical, but in any case until very recently, Rosetta *did* remember in a way similar strings changed in the templates. They used to appear as suggestions, which were nothing else than fuzzy strings. I am suspecting this has to do with the recently approved Rosetta blueprint related to not showing fuzzy strings on the Launchpad UI, but I would have thought that at least the strings would appear in the downloaded PO file. But unfortunately, that's not the case, either. Simply discarding fuzzy translations can lead to lots of wasted work. Just considering a quite usual case: a package 100% translated, with strings where the developer has forgotten to add the period at the end of the sentence. If he later releases a new template where he simply adds this period on the strings where it was missing, after a usual "refresh" with gettext tools (e.g. intltool) all those strings will appear as fuzzy in the PO file, which the translator can then translate quite quickly and without need of duplicate work. Now if all those fuzzy strings had just been supressed (imagine just a few documentation strings, which are usually quite long), all the previous work cannot be used as a reference and will simply have been wasted. Note that at the moment I am only guessing that the problem lies in the removal of fuzzy strings. Something more knowledgeable on Rosetta might have something else to say here. Regards, David. -- ubuntu-translators mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-translators
