On Tue, 21 May 2013 14:13:24 +0200 Peter Nermander <peter at nermander.se> dijo:
>> Rolf, are you saying that there is no way, >> that present-generation Scribus can handle combining characters >> properly? >From what I understand: No. > >Currently Scribus places glyphs one at a time on the canvas and >Scribus does not support placing one glyph on top of another to >combine them. > >I also think that is in line with unicode: Each glyph used should have >it's own code. How would otherwise things like sorting work? I use combining diacriticals constantly for work in linguistics, where I must enter characters from the IPA. Some fonts have pre-composed glyphs for some of the characters, but no font has pre-composed glyphs for all characters that I need. For those who do not understand what I am talking about here is a simple test: Type an n (in your favorite text editor or in Scribus) and then (assuming you are on Linux) type Ctrl-shift+u and when the cursor changes type 329, and then a space or enter. You should get a diacritic under the n that looks like a short vertical line. (This is called the syllabic mark, and it means that the n is acting as a vowel, as it often does in English, German, and a number of other languages.) Having done that, repeat the exercise with an m. Now you will see that the syllabic diacritic is not centered under the m as it was with the n. For a further exercise, repeat with an l, which will further demonstrate the problem of centering the diacritic. I understand that recent versions of InDesign understand that the text contains a combining diacritic and will automatically center the diacritic. There is no such feature in Scribus or LibreOffice. I once had to do some transcriptions in Scribus where I needed a lot of diacritics and I had to kern each one manually to get the diacritic to line up. After doing a few I made myself a table of what fractional point values I needed in order to kern each different character and its diacritic. It was a PITA. I don't know how InDesign does it, but I imagine one could write a script to accomplish the same thing in Scribus. All you'd need is a table for all possible character combinations and then do a "search and if found, kern," and run the script on the entire text. This would not solve Martin's problem, as he is looking for a way to type them correctly on the fly. I don't even know if I understood his problem completely.
