The Ur-Quan Masters Discussion Forum

I've created a new page in the Ultronomicon for information on translation projects. I advice all those who are or want to be involved with such projects to read it and perhaps provide additions.
(Link (http://uqm.stack.nl/wiki/index.php?title=Translations)).

Sorry for possible offtopic, but I couldn't found the answer in the link you provided.
How to force UQM show symbols with numbers 128-255?
I copied 65.png (latin 'A') to 192.png (russian 'A') to same directory (\comm\comandr\comandr.fon\), saved it to uqm-0.4.0-content.uqm file. And put some 'A' letters in the dialog. But when I talk to starbase commander the text shows incorrectly.

I'm assuming you're using UQM 0.4.

Judging from your example, you're assuming the text is in Windows-1251 while UQM (according to the Ultronomicon) uses UTF-8. As UTF-8 is a variable-length encoding and uses byte values from 128 to 255 for multi-byte characters, putting the byte 192 in a file and expecting to get the character 192 shown in UQM does not work. Convert your text file to UTF-8 and change the file names of the additional characters to correspond to their Unicode (http://www.unicode.org/) values.

Interestingly, the Finnish and German translations both seem to be in ISO-8859-1 (Latin 1), as they are apparently based on UQM 0.3 with the fix for chars above 136.

I don't know what editor uses unicode. FAR, notepad and MS Word can't. I could write a converter, but U0400.pdf that describes cyrillic codes is password protected. It seems that everything is done to make translation process impossible.

I see only way to solve this: change symbols 65-127 to russian. To get rid of every latin symbol in game. That's what I call "Per anus ad astra"...

P.S. I've translated many games from english to russian. Civilization 3 and Warhammer were the largest of them with hundreds kilobytes of texts. Help me with fonts and formats and I promise there will be a russian version of UQM.

Modern versions of Notepad should be able to save files as UTF-8. And there are plenty of other editors and existing conversion tools that may help you. Google is your friend.
http://www.jedit.org/ looks promising.

The .pdf unicode tables from unicode.org are not password protected. If you get a message that "your viewer cannot decrypt the file" you're using a version of Acrobat Reader that's too old. The solution is obvious.

As for "It seems that everything is done to make translation process impossible.", I'm not going to bite.

As you can read on the Ultronomicon page I linked to, a Russian version is already underway. You may want to join that team instead of doing double work.

I like Emacs (http://www.gnu.org/software/emacs/emacs.html) myself, but it's a bit... idiosyncratic. But even Notepad can save in UTF-8 (http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/win_notepad_utf_description.mspx).

I fixed the problem by add few line in the function "getCharFromString"
(src/sc2code/strings/unicode.c)

If by "fixed", you mean "changed UTF-8 handling in such a way that the ISO Latin-1 codes for ä, ö and ü (and the same in upper case) are passed straight through the decoder, causing lots of seemingly random corruption in proper UTF-8 text in a wide variety of languages", I agree. Instead of converting the text files to UTF-8, you hack UQM to make it display your Latin-1 text despite the best efforts of the UQM core team to move to Unicode so that translation projects don't need to modify UQM. You've missed the point of using Unicode completely, just like our Russian friend.

Why does everyone seem to have such a problem with UTF-8? Every editor I use under XP and Linux supports UTF-8 properly, and most of the time, UTF-8 files are even autodetected.

Sorry about the rant; I don't mean to offend the German translation crew (thanks for the source release, BTW; I was curious about how you were handling character set issues). However, I think the translators are increasing their own burden by forking UQM unnecessarily. Translations should be drop-in content replacement packs, not complete rewrites! This makes life a lot easier both for the maintainers of the translations and the users.

I must hack UQM in any case, because to translate the hard-coded text.

How can i convert the files to UTF-8 ?
Which font is a UTF-8 font?

See this posting (http://uqm.stack.nl/forum/index.php?topic=1952.msg26648#msg26648).

True, but this will probably change before 1.0. The long-term goal is to use the same executable for all translations.

In modern Windows versions, Unicode is used internally. Essentially, this means that most fonts are Unicode fonts, although few of them contain the entire character set. The same applies to most of the relevant Unix libraries. Unicode is, loosely speaking, a 21-bit character set as opposed to the usual 8-bit character sets, and supports every known written language short of Klingon.

Unicode can be encoded in several different ways. UTF-32 is a straight-forward but very inefficient way of encoding Unicode (represent each 21-bit value as a 32-bit word). UTF-8 encodes 0 to 127 as single bytes (making all 7-bit ASCII text valid UTF-8) and uses the values 128-255 to indicate multi-byte characters (values of 128 and above). You could consider UTF-8 to be a compressed form of Unicode and UTF-32 to be the "raw" encoding.

Most modern editors can save as UTF-8. In Windows XP Notepad, the character encoding is an option in the "Save As..." dialogue box.

The reason I'm surprised that saving to UTF-8 is a problem is that the Linux distribution I use has been defaulting to UTF-8 for about three years.

GNU recode (http://directory.fsf.org/recode.html) is a convenient character set conversion utility.

Thats it  ;D

I convert the files if  the first UQM with outsourced text is finished.

Hi. I have whole translated text to russia, but collided with some problems with fixed length strings. This is a months abbreviations(FEB, JAN...), string "ON" and "OFF", "Enter name"...and maybe some other.

As far as i know, cyrillic char has a two byte len, and english has one. So, i can translate "FEB" to string with one cyr char and one eng char :)

Do you planning to fix this errors? If yes, ill try to report about other similar problems.


ps. sorry about my english. text translation was made by other peoples ;)

I just fixed the issue with months in cvs. But what's wrong with ON/OFF and "Enter Name"?

Also read this (http://uqm.stack.nl/wiki/index.php?title=The_Ur-Quan_Masters_Technical_FAQ#I_don.27t_want_to_make_a_Bugzilla_account._Is_it_ok_to_let_you_know_about_my_bug_in_another_way.3F).


Edit; Ok, I see the problem with "Enter Name" (an input issue rather than an output issue).
That leaves ON/OFF. What's wrong with it?

Ups, sorry, it was bug with "Cyborg" string (not ON/OFF).
If translated(to cyrillic) 'Cyborg' string len = 6, then game crashed.

Ok.

The Cyborg issue is now fixed in CVS. Inputting multibyte characters will still not work yet as the saved game format would need to be changed, we we won't do until we start using XML.

Do you plan to move all text (from *.txt) to XML? When you hope to finish it?
And what about, 'hard-coded' text, like in game menu, planet info screen, "crew" and "fuel" in ship menu, and other.

All text, and many other game data too. Including the stuff that's now hardcoded.
I can't say anything about when it will be finished. I had it planned for a long time ago, but it's still in the begin stages.