Displaying cyrillic text in FM

Report · Apr 15, 2008

Hi,

I've a problem getting dita topics with russian displayed correctly in frame.

External xml parsing shows no errors and file displays correctly in browsers, but frame (7.2) complains about characters:
XML Parser Messages (Document Instance)
Line: 1 Unknown Char: С
Line: 1 Unknown Char: и
Line: 1 Unknown Char: с

and display is garbage (mostly dots, commas and dashes). If I change the XMLDisplayEncoding to, say, ShiftJIS from UTF-8 in structapps, I get cyrillic characters visible (and no parsing errors), but they look monospaced, every letter and word space looks the same. This is clearly a font mapping issue, but I cannot figure out how?

I've got cyrillic versions of fonts used available and codepage 1251 is installed. Could this be solved with "virtual fonts" in Windows and how? (no experience..)

Thanks in advance,
Martti

Report · Apr 16, 2008

Hi Martti,

FrameMaker 7.2 does not support the UTF-8 XMLDisplayEncoding, this is an error. It knows FrameRoman and the four Asian encodings.

What you observe with FrameRoman is correct: Cyrillic characters are not supported (as Russian etc. is officially not supported by this FrameMaker version).

If you choose one of the Asian codepages, you note that those codepages have support for Cyrillic characters, but for display you can only use appropriate Asian fonts (which most of the time are monospaced). The existence of Western/Cyrillic/Greek characters in Asian fonts is kind of a fall back in case you have no real fonts for those glyphs.

The solution: FrameMaker 8. Or a preprocessing XSL stylesheet that changes the Cyrillic characters into the positions of Western characters which then appear correctly when used with a virtual font like Arial CYR.

I have implemented this several times, but it is always a custom implementation because one has to look how to detect the current language/codepage and/or if attribute values have to be translated as well.

I don't know if my email appears somewhere, just check my web site http://cap-studio.de/ at "Kontakt".

- Michael Müller-Hillebrand

Report · Apr 16, 2008

Thanks Michael,

Well, it looks like another stylesheet with some character mapping for cdata and attr values..is there any mapping table, I could use Times New Roman CYR font to do the job. Source xml provides the lang info.

Martti

Report · Apr 16, 2008

You may want to download the FM 8 or TechCommSuite eval and see if the unicode fix makes life easier.

Art

Report · Apr 16, 2008

Hi Art,

FM 8 is viable solution if it doesn't introduce any weird "functionality" in xml roundtrip and using existing templates, edd's and rw-rules and especially FDK applications. Any experiences when going from 7.x to 8?

Martti