Re: [Translation-i18n] distribution coverage stats or how to keep motivation

[Clytie S. <cl...@ri...>]

Regarding TM...

On 24/02/2005, at 10:56 PM, Dwayne Bailey wrote:
>
> No they would be different things.
>
> TM is good for sharing already translated strings and will, if
> implemented well decrease the amount of effort.  Although I have
> generally found only about a 10% match.

It depends on the type of file. If you create fairly specialized=20
compendia, you can get a better match, especially if you tolerate some=20=

degree of fuzziness. I've found that various "x"-utils files have a lot=20=

of strings in common, while the iso-"x" files are a completely=20
different lexical area. A partial match can also save a lot of typing.
>
> What I am more concerned with are those strings that are visible as
> apposed to those that report some corner case error.

Yes, we need priority based on user-interaction.

>  So a unique entry
> on the menu eg "Paste Quotation" (found that on Evolution) vs "You
> bladget exploded while forward seeking through the peripheral maze of
> life" which nobody has ever seen appear in a running version of the
> program.

Sounds like the developer comments in Gaim, where they use those fields=20=

to snipe at each other! :D  There's also a string about mating=20
butterflies...

from Clytie (vi-VN, team/nh=C3=B3m Gnome-vi)

Clytie Siddall--Renmark, in the Riverland of South Australia

=E1=BB=9E th=C3=A0nh ph=E1=BB=91 Renmark, t=E1=BA=A1i mi=E1=BB=81n s=C3=B4=
ng c=E1=BB=A7a Nam =C3=9Ac=