Re: String freeze break requested for Orca for GNOME 2.17.92

We retract this request and will address this and other i18n/l10n
issues for GNOME 2.19/2.20.  As part of this retraction, we are
also pulling the special nautilus script from GNOME 2.18 and will
work to get it in for GNOME 2.19/2.20.



PS - Regarding the question of concatenation of sentences: what
     we do in many cases is concatenate phrases.  Since we're
     dealing with speech synthesis, our goal is to present the
     most important information first - so we'll present the
     most important phrase first followed by other phrases.
     We also need to take care with how we use punctuation as
     it tends to introduce too much dead space in speech
     synthesis.  So, we'll often concatenate the phrases
     using " ".  While the resulting output does not always
     result in perfect grammar, but it does result in a more
     compelling user experience.

     I realize these goals may be in some tension with people
     who are used to seeing text on a display, and I realize
     we've also blown it in spots.  I look forward to fixing
     this for GNOME 2.20, but I'm going to need some
     understanding that the Orca user is not the typical
     GUI user and we need to present things/strings a little
     differently to them.

Christian Rose wrote:
On 2/22/07, Willie Walker <William Walker sun com> wrote:
Hi All:

Attached is a patch for the ngettext change as requested in response to
the original request.  In discussing this issue with Andre Klapper on
IRC, we agreed the move forward plan is this:

1) Limit the break request to  The feature supported by it
comes at the request of our users and they really want to see it get in
for GNOME 2.18.

2) Postpone the changes to GNOME 2.19.  My feeling is that
there are too many strings to mark as translatable and I'd rather not
put undo burden on the l10n team for GNOME 2.18.

All we need is one thumbs up on this.  :-)  If there is sufficient
pushback, however, we will pull the specialized nautilus support from
GNOME 2.18 and wait until GNOME 2.19.  But...we just need to know one
way or the other.

So, I take it that the change you want is this one:


                        itemCountString = _(" %d items") % itemCount

I think there are two issues with this change itself:

A) This message really should use ngettext, otherwise this message
will be broken in any locale that uses other rules for plural than
English. See
So something along the lines of

                      itemCountString = ngettext(" %d item", " %d
items", %itemCount) % itemCount

(I don't speak Python, but you probably get the idea)

B) Why is there a need for a space at the beginning of the message?
When developers add spaces at the beginning of strings, it always
makes me suspect that there's some other problem with the string as
If the space is there for padding or concatenation purposes, then
that's broken localization-wise too. Using spaces for padding is
broken for RTL locales.
If the space is there for concatenation purposes, then that's broken
for all localization efforts, since proper localization of sentence
fragments out-of-context is not possible.

Furthermore, this change is suggested very late in the cycle and in
the string freeze. I feel that either solution to these problems is
not entirely adequate, since there appears to be several issues at
hand with this string.

I hence feel reluctant to approving this this late in the cycle. It's
better to adress this properly later. However, Danilo might be of
another opinion, and in that case I trust his judgement.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]