Gentoo Archives: gentoo-desktop-research

From: Alastair Tse <liquidx@g.o>
To: gentoo-desktop-research@l.g.o
Subject: Re: [gentoo-desktop-research] Report of the desktop-research meeting.
Date: Thu, 22 Jan 2004 00:13:45
In Reply to: Re: [gentoo-desktop-research] Report of the desktop-research meeting. by foser
Hash: SHA1

On 21 Jan 2004, at 22:58, foser wrote:

> On Wed, 2004-01-21 at 21:01, Donnie Berkholz wrote: > >> Since avenj was too lazy to subscribe =P : >> <avenj> wasn't there some kind of potential performance hit there? >> <avenj> i thought i was thinking someone thought part of the issue >> with >> RH's performance was utf8 default, but now i'm thinking i'm not sure > > You don't mean python's alleged performance hit (this got shortly > discussed on -dev by lqx) ? To attribute a full systems performance to > UTF-8 sounds far fetched to me. Anyway, the benefits likely outclass > performance loss.
Exactly, the performance hit is minimal for people who don't care about UTF-8. The only performance hit was in python _if_ you used UTF-8 in a script/application. And the performance hit was not in processing time, but in memory consumption. UTF-8 is pretty much essential to adoption by middle-eastern, european and cjk languages. It could probably be taken on as a desktop project, but I'd much rather it be a Gentoo-wide thing, because it affects much more than the desktop (for instance, if you password has an accent in it, then it will work differently under UTF-8.) BTW, in respect to avenj's comment, there is no noticable performance hit on Redhat 8/9 because of UTF-8 support, unless you mean the fact that it would make glibc compiles a little bit longer because it generates more locales with UTF-8 character sets. Cheers, - -- Alastair 'liquidx' Tse >> Gentoo Developer >> | -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin) iD8DBQFADxW3OM4cezkHFPYRAgZFAKCf3GtGXmhFoXzxxaCuavuS+xiJHgCgp2QG RgZeERo8hyWD4/qC4e4GJ1U= =rpT+ -----END PGP SIGNATURE----- -- gentoo-desktop-research@g.o mailing list