GNOME Bugzilla – Bug 774377
Use Unicode in translatable strings
Last modified: 2016-11-25 09:55:42 UTC
Created attachment 339773 [details] [review] Use Unicode in translatable strings Attached patch converts ASCII characters to Unicode, as recommended by <https://developer.gnome.org/hig/stable/typography.html>.
Thanks for a bug report, though, ehm, well, you should ask before doing any such large change. Personally, I dislike "compressed three-dots", they distract the whole view visually for me. There had been a bug report for this filled some time ago already. The unicode letters can be a problem in the XML files, similarly as for example the '<>"' letters are. I do not say they always are, just that they can. I'm not going to commit this right now, neither the other bug reports about the same (bug #774331, bug #774364, bug #774365).
(In reply to Milan Crha from comment #1) > Personally, I dislike "compressed three-dots", they > distract the whole view visually for me. There had been a bug report for > this filled some time ago already. The Unicode ellipsis looks almost identical to three dots in proportional fonts that are used in UIs, but has certain advantages. For example, it won't get moved to another line like this.. . > The unicode letters can be a problem in > the XML files, similarly as for example the '<>"' letters are. I do not say > they always are, just that they can. > How does that work with translations? They are already full of Unicode characters.
(In reply to Piotr Drąg from comment #2) > The Unicode ellipsis looks almost identical to three dots in proportional > fonts that are used in UIs, In the time I saw it it didn't (couple years back). That's why I dislike it. > but has certain advantages. For example, it > won't get moved to another line like this.. > . I see. > How does that work with translations? They are already full of Unicode > characters. There is a difference between XML "source" file and the localized strings. The localization happens (usually) only after the string is read from the XML file, which makes it easier for translators. I've got an idea yesterday, but didn't write it here: the HIG doesn't say to whom it recommends the usage of Unicode characters. I mean, it doesn't matter what the source code has, what matters is the localization itself. The applications are run in a C locale (thus with strings used in the code) very rarely, the same as for example different languages can have different rules for double quotes, like my mother language uses „” instead of “”. As I can have even written in the code "bla bla" and the translators can translate it into something meaningful, like "En error happen while reading the message", then the same applies to the Unicode letters usage. That is, it will be better to educate translators and not to clutter the code with Unicode characters, which in turn would add more work to translators right now anyway, due to so many changed strings. I can open a thread about this suggestion on a mailing list of your option, if you prefer.
*** Bug 774365 has been marked as a duplicate of this bug. ***
*** Bug 774364 has been marked as a duplicate of this bug. ***
*** Bug 774331 has been marked as a duplicate of this bug. ***
I believe the HIG recommendations are meant for the English interface. There is no English (en_US) translation, it uses strings from the source code. I have already talked to translators about this change and encouraged them to implement their languages' typography rules: https://mail.gnome.org/archives/gnome-i18n/2016-November/msg00033.html Also, I really don't think “” are a problem in XML - they are rather certainly not special characters like "".
(In reply to Piotr Drąg from comment #7) > I believe the HIG recommendations are meant for the English interface. There > is no English (en_US) translation, it uses strings from the source code. Aha, interesting, I thought (and only thought) that the en_US has its own translation too. My fault. > I have already talked to translators... Nice. > Also, I really don't think “” are a problem in XML - they are rather > certainly not special characters like "". You are right, "unicoded" double quotes are not a problem, the libxml should not treat them as a special character.
In the light of your corrections of my wrong thinking, I guess I can commit your changes into the projects. I'll only remove the change with the triple-dot. Even they look almost identical with a proportional font, they look horrible with a monospace font. My editor uses monospace font, that would be just killing me :) I also cannot assure that each place with the triple-dot will always use a proportional font, thus let's have it as a support for the idea of not using it (if you'd like to know what I'm talking about, then try yourself, or simply open the patch in a web browser (Firefox here) and search for the triple-dot; the patch is shown in the monospace font in Forefox for me).
Also see bug 774331 comment #4. Created commit bc5f327 in evo master (3.23.3+)
I wish you would reconsider using proper Unicode ellipses, as GNOME in general is moving to using them in all user interfaces, plus I assure you it's a matter of getting used to - I've been there. :) Nevertheless, what you accepted is a huge step in the right direction and I'm really happy with that!
(In reply to Piotr Drąg from comment #11) > I wish you would reconsider using proper Unicode ellipses, as GNOME in > general is moving to using them in all user interfaces, plus I assure you > it's a matter of getting used to - I've been there. :) Yeah, my opinion tends to change during the time, thus it's possible I'll reconsider it in the future. How far future I cannot tell at the moment. :)