[Sparkle] Incomplete localizations

[Jerry Krinock]

On 2007 Aug, 01, at 9:03, Justin Bur wrote:

> I wonder why so many incomplete localizations are present in the
> Sparkle bundle.

I would say, because this is real life.

> As far as I know, there is no value in shipping a
> localization bundle that simply contains a bunch of English strings.

Yes, but if most of the strings, and all of the commonly-seen  
strings, are translated, there should be some value in that, although  
I've heard contrary opinions on this.

> The system has its rules for dealing with missing localizations if
> the current app's language is not available.

Yes, it uses the next preferred localization.

> Moreover, incomplete
> localizations are subject to bitrot (not a huge concern thus far with
> Sparkle, but it will become a problem if the project undergoes
> significant change).

Well, if they are not maintained, they are certainly subject to  
becoming more incomplete as new strings are added.  But the  
translated strings should not become incorrect, unless the programmer  
changes the context of a given key, which is a programming error.

> Thus I propose: all localizations that have not been fully translated
> and verified by a native speaker should be removed.

That's a wonderful but lofty goal, especially if you imply that the  
translation and verification are to be done by two different people.   
That would be a great achievement in any software project!

Finally, please tell us how you concluded that there are "incomplete  
localizations", so we can track this down.  I saw a case recently  
where the localization .strings file was complete, but the code was  
not using the string.  (Last week, Andy committed a patch to fix that  
one.)