On Wed, Dec 07, 2011 at 05:21:50AM -0600, Jim Warner wrote: > Your approach to nls strings conflicts with the gettext documentation and > common sense. You've intentionally broken up large and/or multi-line strings > into single words or lines. The documentation suggests strings should be > combined into paragraphs wherever possible so that translators don't lose > important context. I have had experience in doing this stuff before and yes it is essential that all of the string, even if it is multi-line, is in one chunk. It just plain won't work otherwise. > Additionally, breaking 'headers' into individual words practically guarantees > the translation will result in ugly, misaligned output. Those are the very > strings that need not only context but some 'Translation' notes as well. > Except for top, no other program provides any translator guidance whatsoever. Headers are tricky ones. The problem is you do want them all to be the same, but there is no guarantee the words are the same length. I've got this: #: src/fuser.c:1186 #, c-format msgid "%*s USER PID ACCESS COMMAND\n" msgstr "%*s UTENTE PID ACCESSO COMANDO\n" I'm a little worried that things won't line up correctly here. Actually checking the program it doesn't for english anyhow, hmm. NLS also brings with it a longer release cycle as you need to freeze the code, send the updates, then release. - Craig -- Craig Small VK2XLZ http://enc.com.au/ csmall at : enc.com.au Debian GNU/Linux http://www.debian.org/ csmall at : debian.org GPG fingerprint: 5D2F B320 B825 D939 04D2 0519 3938 F96B DF50 FEA5