i18n: Avoid sentence puzzles

Hi,

in order to enable translators to prepare proper translations, sentence puzzles have to be avoided. While it makes perfect sense for English, some languages may have to separate those words to sound or even be correct.

The attached patch demonstrates a change to achive that. I did not test it because its purpose is only to raise awareness and start a discussion about this topic. After all the question is, how important translations are for a tool like Git. I have started a German translation but many things are really hard to translate.
1) Many words are used in a germanised way, so translating them not only feels awkward in some cases but it also might confuse users who are used to the original wording.
2) English is a language that can be used in a very compact way. In German that can feel dumb or even rude. So texts can grow up to twice or thrice their size. That can clutter the terminal appearance quite a bit.

Given those problems many people avoid using command line tools in their language and I see the usefulness of translations rather limited.

So my question would be: Is it considered worth it to extend the code for translators' and translations' sake? If so, I would be glad to help with that.

Regards

From 7b1475cbffe120fdae1b46a2974a7b94846702c4 Mon Sep 17 00:00:00 2001
From: Frederik Schwarzer <[email protected]>
Date: Mon, 30 Jan 2012 12:02:46 +0100
Subject: [PATCH] Avoid puzzle sentences.

---
 builtin/branch.c |   27 ++++++++++++++++++---------
 1 files changed, 18 insertions(+), 9 deletions(-)

index 7095718..a07ac54 100644
@@ -154,18 +154,17 @@ static int delete_branches(int argc, const char **argv, int force, int kinds)
 	const char *fmt, *remote;
 	int i;
 	int ret = 0;
+	int is_remote = 0;
 	struct strbuf bname = STRBUF_INIT;
 
 	switch (kinds) {
 	case REF_REMOTE_BRANCH:
 		fmt = "refs/remotes/%s";
-		/* TRANSLATORS: This is "remote " in "remote branch '%s' not found" */
-		remote = _("remote ");
+		is_remote = 1;
 		force = 1;
 		break;
 	case REF_LOCAL_BRANCH:
 		fmt = "refs/heads/%s";
-		remote = "";
 		break;
 	default:
 		die(_("cannot use -a with -d"));

@@ -189,8 +188,11 @@ static int delete_branches(int argc, const char **argv, int force, int kinds)
 
 		name = xstrdup(mkpath(fmt, bname.buf));
 		if (read_ref(name, sha1)) {
-			error(_("%sbranch '%s' not found."),
-					remote, bname.buf);
+			if (is_remote) {
+				error(_("remote branch '%s' not found."), bname.buf);
+			} else {
+				error(_("branch '%s' not found."), bname.buf);
+			}
 			ret = 1;
 			continue;
 		}

@@ -211,14 +213,21 @@ static int delete_branches(int argc, const char **argv, int force, int kinds)
 		}
 
 		if (delete_ref(name, sha1, 0)) {
-			error(_("Error deleting %sbranch '%s'"), remote,
-			      bname.buf);
+			if (is_remote) {
+				error(_("Error deleting remote branch '%s'"), bname.buf);
+			} else {
+				error(_("Error deleting branch '%s'"), bname.buf);
+			}
 			ret = 1;
 		} else {
 			struct strbuf buf = STRBUF_INIT;
-			printf(_("Deleted %sbranch %s (was %s).\n"), remote,
-			       bname.buf,
+			if (is_remote) {
+				printf(_("Deleted remote branch %s (was %s).\n"), bname.buf,
 			       find_unique_abbrev(sha1, DEFAULT_ABBREV));
+			} else {
+				printf(_("Deleted branch %s (was %s).\n"), bname.buf,
+			       find_unique_abbrev(sha1, DEFAULT_ABBREV));
+			}
 			strbuf_addf(&buf, "branch.%s", bname.buf);
 			if (git_config_rename_section(buf.buf, NULL) < 0)
 				warning(_("Update of config-file failed"));

1.7.8.3

The sample you supplied for 'branch' vs 'remote branch' is a good change in the logic to choose what translatable message to use, and I do suspect that there are other instances of problems of this kind.

I am expecting that we may see some activity to the logic in that file during this cycle, and we would need to be careful not make people step on other people's toes, though.

I added the i18n support you're having problems with, and I completely agree that this is the sort of thing we need to do.

Unfortunately when I added the i18n support I didn't have time to get rid of all these sentence puzzles (or, to put it another way "lego" sentences). You should never force translators to translate partial sentences, you should always provide them with full sentences that they can translate completely, even if that means that there's some duplication for some languages.

Problem reports like this one are exactly what we need at this point for i18n, we need people spotting these issues, and then we can fix them.

As a translator (though not git's because my time is limited and GUI apps have my priority), I completely agree and support this. There are other places where a sentence is broken down into many short phrases. It's hard for a translator to find a good translation in such cases.

I remember there was also a patch about "1 file vs 2 files" in diff summary, which was rejected because it would break scripts. I think grammar patches should be allowed at least for interactive use (i.e. either pager is on, or std{out,err} is tty).

Duy