Re: English translations stats

[Abel Cheung <deaddog deaddog org>]

> There are several problems with not doing so, i.e. leaving msgstrs
> empty. The most important one is that it becomes technically impossible
> to distinguish between a completely new message that needs translation,
> and an old one that was previously verified to not need any translation.
> Also, leaving msgstrs empty on purpose completely screws up the
> translation statistics.

Like Bastien have said, sometimes there's no differentiation between
en_GB and en_US in some modules. Even when there is some, it's just a
low percentage of strings. I suppose it's lower than 10% -- that means
en_GB is an automatically supported language even when no en_GB
translation exists, if "supported" means "strings are displayed
correctly in localized version". Same for en_CA.

Of course duplicating msgid -> msgstr would make calculation easier
technically, but this puts some seemingly useless workload on en_GB
maintainers. Anyway it's en_GB, so the decision is probably up to
en_GB maintainers.

Abel


> 
> Both of these reason is why gettext has a "msgen" utility nowadays,
> useful when translating into various other flavours of English. It
> copies the unmodified msgid into the msgstr for every message but
> fuzzymarks them at the same time, so that it is absolutely clear, both
> to a human and a tool, what translations have been verified to be
> correct and which ones haven't.
> 
> 
> Christian
> 

-- 
Abel Cheung
Linux counter #256983   | http://counter.li.org
GPG Key: (0xC67186FF)   | http://deaddog.org/gpg.asc
Key fingerprint: 671C C7AE EFB5 110C D6D1  41EE 4152 E1F1 C671 86FF

PGP signature