utf8 vs iso8859-1 speed/responsiveness

[Hideki Hiura]

> From: Pedro Fernandes Macedo <webmaster at margo.bijoux.nom.br>
> Not exactly...  The issue is the input file.
> In RH 9 , the input file probably was iso8859-1 and then it was 
> processed without any conversion , as the whole system was using 
> iso8859-1. Now in fedora  , the input file has to be converted by the 
> app to UTF8  , so this extra step means that FC will be a bit slower 

In general this is true, but in this case, not exactly :-). 

Glibc internal encoding is UTF32/UCS4, and modern toolkits, thus
major desktop apps as well, OOo, all internal encoding is UTF-8, 
on RH9 as well.

Even you set the locale to *.iso8859-1, and keep file contents in
iso8859-1 encoding on RH9, whenever apps open it, it is converted to
UTF-8 for processing upon reading, and convert back to iso8859-1 when
apps saves it, even though it appeared as the whole system was using 
iso8859-1.

This architecture is unchanged between RH9 to FC2, so encoding 
conversion happens everywhere on the fly.

So regardless of RH9 or FC2, and regardless with the locale
you set(*.UTF-8 or *.iso8859-1 or whatever), at least the files 
encoded in UTF-8 takes less for I/O than other encodings.

Regards,
--
hiura@{freestandards.org,OpenI18N.org,li18nux.org,unicode.org,sun.com} 
Chair, OpenI18N.org/The Free Standards Group          http://www.OpenI18N.org
Architect/Sr. Staff Engineer, Sun Microsystems, Inc, USA   eFAX: 509-693-8356