Re: [EWMH] _NET_WM_WINDOW_TYPE_AUXILIARY

[Nicolas Mailhot <nicolas mailhot laposte net>]

Le jeudi 18 octobre 2007 à 19:17 +0000, Tuomo Valkonen a écrit : 
> On 2007-10-18, Nicolas Mailhot <nicolas mailhot laposte net> wrote:
> > You can rave all you want things should be nicely tagged with encodings
> > but they aren't and won't be till an awful lot of otherwise perfectly
> > working code is rewritten.
> 
> The big problem is that _new_ code is being written specifically
> for a monoculture; old code simply didn't care.

Old code did care. It just cared in a miriad incompatible ways which all
more or less assumed the default was what people wanted it to be, and
broke horribly otherwise (I have access to a fun pile of documents that
uses a custom encoding just because there was no 8-bit encoding with the
right mix of symbols, and relies on a one-of-a-kind font with the same
cooked encoding to be user-readable)

> > So the next best thing is a good universal default. Which UTF-8 is. So
> > live with it (or join unicode.org to get it improved).
> 
> That's just another reason why it's pointless to bear with FOSS, as it
> can do nothing better than the commercial OSes. 

FOSS needs to exchange data with other systems (internet remember?).
That means sharing encoding conventions.

> > The single best feature of XML was not making possible to tag stuff with
> > encodings (HTML had it before, as SGML). The single best feature of XML
> > was to select UTF-8 as default encoding, so stuff is internationalised
> > by default.
> 
> Most of the XML files I've seen include a specification of the encoding.

Which is usually UTF-8 because people just use the default (just like
they just used iso-8859-1 in headers because it was the default, and
then stuffed something else inside because iso-8859-1 just did not have
the required encoding coverage). It's a pity the W3C didn't go the full
way and allowed to specify something else – non-UTF-8 XML files win you
nothing and are a constant source of bugs.

> > So it's fun to shot at UTF-8. UTF-8 is ugly. UTF-8 reeks of compromise.
> > But UTF-8 works which was not the case of all the solutions UTF-8 haters
> > dreamed before and still cling to.
> 
> Actually, UTF-8 as an ASCII-compatible mapping from 32-bit numbers to
> 8-bit sequences is beautiful, something that can not be said of most
> other multibyte encodings. However, Unicode or ISO-10646 or whatever
> they want to call the character mapping in the background, is extremely
> ugly.

That's a result of trying to cater to every known human script. Which
needs to be done to digitalise existing stuff. No one so far has proved
it could be done better.

-- 
Nicolas Mailhot