Take the 2-minute tour ×
Ask Ubuntu is a question and answer site for Ubuntu users and developers. It's 100% free, no registration required.

There are certain Romanian characters (ș, î, ț, ă, â) which are handled differently in the different key layouts.

The standard layout makes them look odd, different from the other letters in Libreoffice (and also in MSOfficeXP).

enter image description here

In msoffice 2007 they look the same (that is: all look fine).

In fact these characters are differently constructed in the different layouts, although msoffice 2007 seems to avoid displaying the difference. (Which also has the good effect that its spell-checker has no problems with them; while in LibreOffice the use of these characters with winkeys (and other layouts that display the characters properly) is reported as error (as seen in the image); on the other hand, in msofficexp the spell-checker reports as error any use of these characters with ro-std layout)

The problem is that the characters look bad in LibreOffice, as seen in the image. They look like they used a different font than the letters of the same word.

Can this be helped?


In my keyboard layouts handler (in Xfce) I see 4 possible choices

enter image description here

  • My problem is related to the way this characters have been implemented in the "standard" version of the Romanian layout. The Wikipedia page says that there was a problem with these characters in Windows before Vista and was solved in later MSWin versions. What I am dealing here with is a problem with these characters in the layout that is called the "primary" one on the Wikipedia and "standard" in my panel list.

    enter image description here

  • The "secondary" version (which is "used as the default one by the majority of GNU/Linux distributions", says Wikipedia) is called "cedilla" in my keyboard layouts options; it avoids this problem reported here, but just like winkeys, is not accepted by libreoffice spell check: any special character is seen as error; also, this is a version for programmers: it is like the US English layout, and the special characters are accessed in a way that would not change the usual use for programming (which is English); that is: they are too hidden and hard to get for people that would really use Romanian to write, for these special characters are very often used. As it is said here, the idea is not to change the layout too often for people that rarely use these characters. It looks like so

    enter image description here

  • The Winkeys layout looks like the "standard" (primary),except that the characters 'y' and 'z' switch positions, and some non-letter characters, like quotation marks, have different keys.

  • The ”standard cedilla” layout is identical with ”standard”, and it avoids the problem reported here by entering characters that, while looking good, are rejected by the LibreOffice spell checker.


Summing up --

The question could be reformulated like so:

The only Romanian keyboard layout accepted by the spell check (in LibreOffice, but also in Firefox and so on), the ”standard”, has special characters that does not look as they should, while the spell checker rejects precisely the characters that look good..

share|improve this question
    
If I set all letters to the same font (Arial) I still see the difference from the posted image. when i select the letter it still says Arial, but looks different. also, if i select just the letter and change its font, it changes, but is also different from others with the same font. if it were just a matter of font, the word (for example ”șase”) would not be seen as error by the spellcheck of msofficexp. –  cipricus Mar 27 '13 at 15:34
add comment

1 Answer 1

up vote 3 down vote accepted

Unlike similar issues with Unicode fonts, this is not really a matter of fonts and glyphs (although particular fonts may not provide both variants). The problem stems from the specific history of Romanian orthography, with the end result that the comma and cedilla variants have different code points. A dictionary written for one style of spelling will always see the other style as incorrect.

Interestingly, the Hunspell ro_RO.aff file for Romanian appears to have the needed rules to work with a dictionary that provides both spellings, but the current Hunspell ro_RO.dic file provides only the comma variants.

You can, however, install a cedilla-based dictionary as an extension in LibreOffice.

The extension is available from the Apache OpenOffice site. The most recent version that I found is dict-ro.1.3-cedilla.oxt (or here - also here).

Doubleclick the downloaded file (or right-click and choose Open with LibreOffice) to install. After installing, close any open LibreOffice elements and restart Writer.

The comma-based word was pasted from Omniglot Romanian and the cedilla-based word was typed with the Romanian-WinKeys keyboard. I had LibreOffice 3.5 handy, but I don't see anything in the extension that should be a problem in 4.x.

Romanian text

share|improve this answer
    
i greatly appreciate your answer. i have just installed the dict-ro.1.3 cedilla and now I have solved two problems. First, altough I prefer using LibreOffice in Ubuntu, my documents are often edited in both windows and ubuntu, and having compatibility in spell-checker is important; second, as seen in the images, the cedilla/winkey version of these characters looks better in linux than the comma/omniglot version, which look out of proportion, a problem in itself which is not solvable as far as i understand. –  cipricus Jun 5 '13 at 10:00
    
I'm glad it helped. I enjoy working on the language questions. The proportions and style of the comma depends on the glyphs in the chosen font. –  chaskes Jun 5 '13 at 14:55
    
what does 'glyph' mean? :) –  cipricus Jun 5 '13 at 15:19
1  
I'm sorry for being unclear. Fonts combine typeface (serif, etc) and artistic design. Glyphs are the bit-mapped images that are displayed. A single Unicode code point may vary greatly due to glyph design, but different fonts can be chosen. Romanian history led to distinct code point pairs for the variant letters. So tools are incompatible and a new font won't help. But the look is from the glyphs. A new font will change that, but may be non-free. A complication is that the MS fonts on Linux were made from old versions that embed some pre-Unicode fallback chars that often do not look very good. –  chaskes Jun 5 '13 at 15:58
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.