[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Splitting locales and i18ndata from glibc into some sane packages

On Fri, Oct 06, 2000 at 03:57:58AM -0700, Joey Hess wrote:
> * Pick the top 5 or 10 languages that are spoken by a large percentage of
>   the world/computer users, and put them in one package, which will
>   probably be about 2 mb in size. Put all the other, rarer languages in
>   another package, which I guess will be 18 mb or so in size. It doesn't
>   satisfy everyone, but it makes things easier for a lot of people
>   without inconveniencing anyone more than they are now.

For my /usr/share/locale (after removing kde-i18n-eo, which would have skewed 
the numbers): 
	top 5, by size (de*, fr*, es*, ru*, ja*): 12 MB	
	next 5 (pl*, cs*, pt*, it*, sv*): 9 MB
	rest, by size: 14 MB
Most of the languages that are spoken by a large percentage of the world
(according to the World Almanac) aren't even in /usr/share/locale: Hindi, 
Bengali, Arabic, Malay-Indonisian. The rest: Spanish, English, Portugese, 
Russian, Japenese, German and French are all up there, except en* (252 KB). 

As I suspected, the big 5/10 are fairly dominant size-wise. Everything
proposed so far seems fairly arbitrary, besides breaking it down by language,
which isn't feasible.

How many people have long LANGUAGE's? If there are some fairly consistent
trends (da:sv:de, for example), we should perhaps package those together.
(I would offer my own, but eo:en:de:la:sw is probably so ubiquitous as to
be obvious.)

David Starner - dstarner98@aasaa.ofe.org
http/ftp: dvdeug.dhis.org
And crawling, on the planet's face, some insects called the human race.
Lost in space, lost in time, and meaning.
	-- RHPS

Reply to: