[Kde-accessibility] Quotation marks in german translation

Luciano Montanaro mikelima at cirulla.net
Wed Feb 13 18:18:08 CET 2008

On Wednesday 13 February 2008 17:39:01 Gerrit Sangel wrote:
> Am Mittwoch 13 Februar 2008 schrieb Luciano Montanaro:
> > More than accessibility, the problem could be that iso-8859-1 and
> > iso-8859-15 charsets do not have those characters. This can be a problem
> > if non utf-8 locales are still in use.
> Well, I do not know how this is in Italian, but in German, äöüß is used, so
> it is not ascii compatible anyway. 

Those characters are also in the iso-8859-1 and iso-8859-15 charsets, so they 
are not a problem, even with an 8 bit locale, they will work. The fancy 
quotes “„ are not, and may be a problem.

> And everything nowadays supports UTF-8, 
> and if not, it really should. I (personally) do not care if someone has
> some problems with unicode characters, just because he is too lazy to
> change.

This is not a user issue, it's a distribution issue mainly, and maybe a font 

And I don't know how many distributions are affected, or how many users are.
So maybe everyone is on utf-8 by now, but I still have some doubt about it.
Distributions I use have UTF 8 locales by default. 

> This is a bit different for accessibility, but eventually (if they 
> do not do now) the accessibility applications should support full Unicode
> as well.

> > I'd be interested in possible issues too, especially since we introduced
> > the «» quotation marks in the Italian translations, and I hope these
> > would not be a problem.
> >
> > I'd actually like to use the ellipses (but how do you type them?) and the
> > n-dash when it makes typographical sense.
> You can install scim (www.scim-im.org), which has a Unicode raw input input
> method. You can then just look in the unicode tables (they are all on
> www.unicode.org) and enter the hex unicode value, e.g. 2026 for …

Ugh. That's not a very nice solution. I may want to use the character, but 
typeing it must be relativel mnemonic. I use an US layout with a compose key 
to type extended character, and I hoped there was some composition sequence I 
didn't know. 

> A bit easier (and for everyday use) is to change your keyboard layout
> in /usr/share/X11/xkb/symbols with e.g. U2026 for …  This is a bit
> inconvenient, because you cannot have _every_ unicode char on the keyboard,
> but for most characters, this should be sufficient.
> example for my elippses key:
>     key <AB09>  { [    period,      colon,                U2026,
> division ] };
> Some characters have “normal” names, too, e.g. the en-dash:
>     key <AB10>  { [     minus, underscore,               endash,
> emdash ] };

This is compose minus minus full-stop here.

> If you do not want to look in the unicode charts (even though they are
> mostly quite straightforward), Wikipedia has really good discussions about
> these characters. You can also just copy & paste them, but I find this a
> bit inconvenient. Changing the keyboard layout is the best way, imho.
> > I also had a patch for the Italian .po files to use ←,→ in the menus
> > instead of "Sinistra" and "Destra", which would directly map to key
> > labels, but as well was not sure if it would break on system without the
> > proper fonts.
> Well, the DejaVu fonts are an extension of the Bitstream Vera fonts, and
> they are under the GPL. Most linux distributions use them as there default
> fonts (or the Bitstream fonts, but I do not know why the do not change
> this). I cannot think of any reason why they should not be used. And they
> have glyphs for _every_ of these characters.

I know. I am quite fond of DejaVu myself. But can we rely on them being the 
default fonts? Actually, font substitution seem to work well, and the arrows 
seemed to show up for me with most fonts anyway. It would be nice to *know* 
if there are problems, however, instead of just hoping all would be well.

> I personally am a bit fed up that the overall development of the font
> display in the computing area is so slow, just because some people are too
> lazy to upgrade there system, without any real reason. Unicode is so old
> and still not everywhere supported, because the people still use iso-8859-1
> or so. They do need a gentle push ;)

The problem is that we do not have control of the complete system, and we must 
be a bit conservative when it comes to these issues.

I think we can have more leeway with the documentation, though -- it's 
converted to xhtml, and I think we can rely on utf-8 there.


More information about the kde-accessibility mailing list