Gentlemen, how could it be that for non-English languages, e.g.,
$ GET http://zh.wikipedia.org/wiki/洪朝貴 |grep '<title>'
still gives
<title>洪朝貴 - Wikipedia</title>
and not
<title>洪朝貴 - 維基百科</title> ?
I.e., how could you forget to translate the most important piece of
the page, the title (suffix)?
(Not the same as Bug 5638, nor Bug 8873.)
Version: unspecified
Severity: normal
URL: http://zh.wikipedia.org/wiki/洪朝貴
The relevant message is 'pagetitle' and it doesn't need any translation. Sitename of zh.wikipedia.org is set to Wikipedia and that's what it displays.
OK, thanks. Exploring
http://ru.wikipedia.org/
http://zh.wikipedia.org/
http://ar.wikipedia.org/
we see only zh has the problem.
Note that in a default setup, $wgSitename would show up here:
$ GET http://zh.wikipedia.org/ |grep siteSub
<h3 id="siteSub">维基百科，自由的百科全书</h3>
but each *.wikipedia.org apparently has a motto replacing it,
so unless one checks the <title> e.g., in one's browser tags, one
might not notice it.
Checking the expansion of {{SITENAME}} proves
that zh.wikipedia.org's $wgSitename is miscofigured, still using English.
Please check the other languages too. I only checked the above three.
The answer turns out to be...
"兩名有簡繁歧義，無法用轉換系統作出變換。--Shinjiman ⇔ ♨ 2007年12月14日"*
The simplified/traditional Chinese people can agree, and it can't be changed by the conversion system.
Surprise surprise. One character ruins all the fun.
Or the converter system should be beefed up.
If it can change the bodies, it should also be able to change the <titles>.
Changing summary and reopening.