Re: More info about #228486 / #235759
On Wed, Jul 28, 2004 at 09:46:40PM +0200, Thiemo Seufer wrote:
> Jens Nachtigall wrote:
> > > a) Use english doublequotes
> > > - the opening quote sign should be subscripted
> > > - opening/closing are identical, this makes nested quotes hard to
> > > read
> > >
> > > b) Use guillemots
> > > - guillemots are very unusual in Germany (In printing I remember
> > > only a few occurences in ~1900 vintage books, and none in
> > > contemporary literature)
> > That is by no means right. I don't know what kind of books you read, but
> > if I take a random look at my library, I get these books with »« :
> > * »Die Hebammen-Sprechstunde« (book about getting children) -> 2002
> > * »Perspektivenübernahme und soziales Handeln« (social studies) -> 1982
> > * »Die Lust am Schauen« (social studies) -> 1986
> > * »Zope« (computer science) -> 2004
> > There are also some books with U201E and U201C, however, these are the
> > minority.
> Well, as said, my impression is a different one, especially WRT periodic
I am not a German native speaker and have thus no opinion, but I was
quite surprised to read that guillemots are uncommon in German, since
most GNU and GNOME translated applications use them. OTOH KDE uses
English quotes, and there is no occurence of ,," in German localized
catalogs, so your conclusion that (b) is worse than (c) cannot be true.
Let's review other points in your previous post:
] - The usual de-latin1-nodeadkeys keyboard layout hasn't even a
] definition for quillemots, it can only be typed in via
] AltGr+<Codepoint>. The X11 keyboard has a common definition for
] all latin charsets (M-y, M-x), but this isn't marked on the
] keyboard, so most people don't know how to type guillemots (and
] they don't have to).
This is irrelevant, this issue is about transliteration, not how text
should be typed.
And even if it was relevant, the same argument works for Unicode quotes
and for French keyboards (some people also argue that guillemots and
accents on uppercase letters are not valid in French because our
keyboards have no such symbols, but those people are usually not very
fond of l10n -- no offense in mind)
] - guillemots are already in use in swiss german, but with _swapped_
] semantics. This will make a confusing reading for swiss people.
Writing a different transliteration for de_CH is trivial, I do not want
to waste my time doing this unless it has a chance to get in.
Note also that de_CH folks already have to parse many guillemots in GNU
and GNOME applications, so they should not be that confused by this
] The last drawback for any change is that it will AFAICS be a long-term
] deviation from upstream glibc. I believe upstream had already made the
] decision to stay with c) until utf-8 is found everywhere.
Why is a deviation a problem *in this particular case*?
[Back to your current post]
> > Regardless from personal impressions of what's the most often used case,
> > Duden (the German authority for orthography) recognizes both as valid.
> > > So a) is IMHO the best.
> > This would be a consense until there is no special Swiss translation.
> I have to raise another point:
> If guillemots are actually in such wide use, we should IMHO avoid to map
> german quotes to guillemots. They may be used in different contexts at
> the same time, or for nested quotes.
In your previous post you asserted that guillemots are uncommon and will
confuse readers, and now you tell that if they are common, they should
not be used. What about English quotes? They are the best choice
because they are quite common and thus will not confuse readers, and
will not interfere with other English quotes since they are seldom used?
> > The big problem is that ,, and " is the something absolutely never being
> > used. If U201E and U201C is not available, people use »...« or "...".
> > Nobody would ever think of replacing U201E and U201C with ,, and ". Do
> > you agree (I only ask to let gotom here your opinion on this)?
> Well, I, personally, use "...", and I think it's the most legible style.
> The human eye/brain does pattern matching, so the correct shape is more
> important than orientation ('ä' vs. 'a:' works surprisingly well),
> which in turn is more important than the position (subscript vs.
It is not clear to me whether you want to have German translators use
English quotes or want to have U201E/U201C transliterated into these
In the latter case, this has been discussed on debian-l10n-german, and
they chose guillemots. My concern is to push what German translators
decided, I do not care whether this is in favor of English quotes or
> > > The last drawback for any change is that it will AFAICS be a
> > > long-term deviation from upstream glibc. I believe upstream had
> > > already made the decision to stay with c) until utf-8 is found
> > > everywhere.
> > you "believe"? Can you give any pointers? I think, upstream will also
> > change this as soon as someone knowing the German language and its
> > habits will have a look at it.
> The upstream lead developer is german, he wrote the original
No, he gets it from some specs (I do not remember exactly, maybe in
Unicode) which give transliteration for some symbols. But these rules
are given in general case, and are superseded by national rules when
they are defined.
> and he's well known to have strong opinions. I think it is unlikely
> that he will change his mind.